Document Rj6j1NdkjXEYR8JdbVvbLDNJ7

JOEM Volume 45, Number 10, October 2003 1059 Cancer Incidence Among Union Carpenters in New Jersey John Dement, PhD Lisa Pompeii, PhD Isaac M. Lipkus, PhD Gregory P. Samsa, PhD A cohort of 13,354 male union carpenters in New Jersey was linked to cancer registry data to investigate cancer incidence during 1979 through 2000. Surveillance, Epidemiology and End Results data were used to calculate standardized incidence ratios (SIRs). A total of 592 incident cancers were observed among this cohort (SIR = 1.07), which was not statistically in excess. However, significant excesses were observed for cancers of the digestive system and peritoneum (SIR = 1.24) and the respiratory system (SIR = 1.52). Workers in the union more than 30 years were at significant risk for cancers of the digestive organs and peritoneum (SIR = 3.98), rectum (SIR = 4.85), trachea, bronchus, and lung (SIR = 4.56), and other parts of the respiratory system (SIR = 11.00). Testicular cancer was significantly in excess (SIR = 2.48) in analyses that lagged results 15 years from initial union membership. Additional etiologic research is needed to evaluate possible occupational and nonoccupational risk factors for testicular cancer. (J Occup Environ Med. 2003;45:1059-1067) From the Division of Occupational and Environmental Medicine (Dr Dement, Dr Pompeii), Division of Psychiatry (Dr Lipkus), and Department of Biostatistics and Informatics (Dr Samsa), Duke University Medical Center, Durham, North Carolina. Grant Sponsor: National Cancer Institute grants CA63782 and CA72099. Address correspondence to: John M. Dement, PhD, CIH, Associate Professor, Division of Occupational & Environmental Medicine, Department of Community & Family Medicine, Duke University Medical Center, Box 3834, Durham, NC 27710; E-mail address: john.dement@duke.edu. Copyright by American College of Occupational and Environmental Medicine DOI: 10.1097/01.jom.0000085892.01486.6a n 2000, 1.2 million carpenters were employed in the United States.1 Car penters perform numerous construc tion-related tasks that may place them at risk of exposure to a variety of known and suspected carcinogens, such as asbestos, silica, wood dust, man-made mineral fibers, solvents, welding fumes, and various metals.2,3 Some cancer mortality studies, which are summarized in Table 1, suggest that carpenters and workers in related occupations are at an in creased risk for cancers that correlate with these types of exposures. Higher-than-expected mortality for can cers of the lung,4-7 pleura,5,8 stom ach,9-11 and buccal cavity and pharynx10 among carpenters and re lated crafts have been observed. Mortality studies provide good es timates of cancer incidence for sites such as the lung and stomach, where the case fatality rate is high (ie, 5-year survival rates <20%).12 How ever, studies of mortality are not useful for certain sites, such as the urinary bladder, breast, and prostate, where 5-year survival rates exceed 75% and 10-year survival rates ex ceed 60%.12 Of interest for carpen ters and wood dust-exposed popula tions are nasopharyngeal and sinonasal cancers, for which numer ous epidemiologic studies have dem onstrated an excess risk, especially among populations exposed to dusts from hard woods.13-17 However, the sinonasal cancers are relatively rare and are associated with low case fatality rates (5-year survival >60%)12; therefore, cohort mortality studies have limited statistical power to detect elevated risks unless the Jicine. Unauth 1060 Cancer Incidence Among Carpenters Dement et al TABLE 1 Previous Epidemiologic Studies of Cancer Mortality Among Carpenters and Construction Laborers All Cancers Bronchus and Lung Pleura/ Other Parts of Respiratory System Rectum Stomach Author Year (n) Hall et al**1*991 (10,824) Hrubec et alt 1992 (248,046) Milham* 1974 (16,443) Milham 1983 (16,843) Registrar11 1986 (3,771) Robinson et al1 1995 (61,682) Robinson et al** 1996 (26,435) Stern et altt 1995 (11,685) Dement et al# 2001 (43,180) n RR n RR n RR n RR n RR - - 152 1.24a 17 3.06a - - 13 1.72b 383 0.9 79 1.0 - - 18 1.2 32 1.60c 3,629 0.97 1,218 1.06b 6 1.33 110 0.85 218 1.07 2,974 1.04b 261 1.18a 2 0.69 NS NS 285 1.20a 3,771 1.05a - - 47 3.61a - - - - 3,731 1.02 1,489 1.16a 14 1.63 - - - - 7,337 1.15a 2,648 1.23a 27 3.01a - - 252 1.30a 3,285 1.00 1,208 1.06b 12 1.13 62 1.20 170 1.44a 2,140 1.00 914 1.17b 8 1.15 26 0.88 55 1.06 a P < 0.01; b P < 0.05; c P < 0.10; NS = not significant, specific ratio not reported. SMR, standardized mortality ratio; PMR, proportionate mortality ratio, PCMR, proportionate cancer mortality ratio; PCIR, proportionate cancer incidence ratio. * PCIR cancer cases among construction workers in New Jersey (white males only included in this table) 1979-1984. Expected deaths based on New Jersey cancer rates. t SMR deaths among US veterans by occupation-known smoking status 1954-1980. Expected deaths based on US death rates. * SMR death claims among carpenter union members 1969-1970. Expected deaths based on U.S. rates for 1968. PMR death certificates among carpenters in Washington State 1950-1979. Expected deaths based on WA mortality. || PMR death certificates among carpenters 1979-1983. Expected deaths based on England and Wales population mortality. H PMR death certificates among carpenters from 19 U.S. 1984-1986. Expected deaths based on US death rates. ** PMR deaths among white, male carpenter union members from 1987-1990. Expected deaths based on US death rates. tt PMR deaths among construction workers in the LIUNA, 1985-1988. Expected deaths based on US death rates. * PMR deaths among NC construction workers, 1988-1997. Expected deaths based on NC death rates. cohort is prohibitively large. Studies of cancer morbidity using casecontrol study designs with cases identified from clinical records or cancer registries have been most use ful in elucidating the risk of sinonasal cancers among carpenters and other wood dust-exposed workers. For example, nasopharyngeal cancer was significantly elevated (PCIR 227; P < 0.05) among construction workers, including carpenters, in a study that used incidence data from the New Jersey State Cancer Registry.18 The purpose of this study was to estimate the risk of all cancers and site-specific cancers among mem bers of United Brotherhood of Car penters and Joiners of America (UBC) in New Jersey by individually matching them with New Jersey State Cancer Registry (NJSCR) data. Cancer registry data are valuable for providing information about incident cases of cancer; however, no infor mation is available through the Reg istry with regard to occupation or workplace exposures. By combining these data sets, we sought to examine the incidence of various cancers, es pecially those with low fatality rates that may not have been detected in previous mortality studies, among a large cohort of carpenters within the same geographical region. We calcu lated standardized incidence ratios (SIRs) for various cancers using na tional incidence rates provided by the Surveillance, Epidemiology and End Results (SEER) program. The SEER program of the National Can cer Institute collects information about persons diagnosed with inci dent cases of cancer from 11 popu lation-based cancer registries, cover ing approximately 14% of the US population, and is considered the gold standard for providing national estimates of cancer incidence. Materials and Methods The current analyses of cancer in cidence among carpenters were un dertaken as part of an ongoing com munication intervention trial to increase colorectal cancer screening among potentially higher-risk bluecollar workers funded by the Na tional Cancer Institute. This study is being conducted in collaboration with the New Jersey Carpenters Health Fund. The New Jersey Car penters Fund provides and adminis ters health and welfare benefits for UBC members in New Jersey. The initial data file used to define the cancer incidence study cohort consisted of 15,375 male UBC mem bers in New Jersey who had partici pated in the New Jersey Carpenters Fund at some time prior to July 1, 2000. This cohort was matched to the New Jersey Carpenters Pension Fund to define starting and ending dates of UBC membership (month and year), pension fund contributions by year, and state and date of last known address for retired UBC members. The cohort was further restricted to 13,354 workers with some period of iicine. Unauth JOEM Volume 45, Number 10, October 2003 1061 active UBC work after January 1, 1979, the starting date for the NJSCR. The UBC study cohort of 13,354 carpenters was matched to the NJSCR through December 31, 2000 using worker social security number as the primary matching criteria fol lowed by last name, date of birth, and gender. The NJSCR is a popula tion-based incidence registry and in cludes all cancer cases diagnosed among New Jersey residents since October 1, 1978. New Jersey regula tions require reporting of all newly diagnosed cancer cases to the NJSCR within 3 months of hospital discharge or 6 months of diagnosis, whichever occurs first. All primary malignant and in situ neoplasms are reportable, except for certain carci nomas of the skin (ICD-9 Code 173). The NJSCR collects demographic data on each cancer patient (social security number, gender, race, date of birth) as well as information on each cancer (diagnosis date, ana tomic site, histology, behavior, and summary stage). Annual follow-up status (alive/dead), dates of death, and underlying cause of death are recorded for each reported cancer case. Using data on primary cancer site (ICD-O-2), histology, and be havior, an ICD-9 code was assigned to each reported cancer using the software program IARCtools (Ver sion 2.01) developed by the Interna tional Agency for Research on Cancer.19 The PC version of Life Table Analysis System (LTAS) (LTAS Version 1.0 days) developed by the National Institute for Occupational Safety and Health was used to calcu late SIRs of all cancers and sitespecific cancers.20,21 The LTAS stratified person-years at risk (PYAR) for each worker into strata by 5-year age groups, 5-year calen dar time periods, length of UBC membership, and time since initial UBC membership. The PYAR for each worker began accumulating on January 1, 1979 or the date of first participation in the pension fund, which ever occurred later. The PYAR for each worker was stopped on the last date of New Jersey union pension fund contribution or last date of known residence in New Jersey for retired workers. Workers were assumed to have been a resident of New Jersey and therefore subject to cancer reporting requirements for each year in which contributions were made to the New Jersey pen sion fund. Only cancer cases occur ring within the periods of PYAR accumulation were used for the SIR analyses. PYAR in each age and calendar time strata were multiplied by corresponding cancer incidence rates from the SEER Program of the National Cancer Institute to calculate expected malignant neoplasms by site. Although race data were not available for the entire cohort, race data were available for cancer cases reported to NJSCR and for workers participating in the colorectal cancer risk communication intervention study. Among 1236 workers with cancers reported to the NJSCR, 97.4% were white; therefore, SEER incidence rates for white males were used for all analyses. SIRs by cancer site were calcu lated as the ratio of observed to expected cases. To account for occu pational cancer induction periods, additional analyses were undertaken which lagged duration of carpenter union work by 15 years. These anal yses discounted person-years and cancer cases prior to 15 years since initiation of work in the carpenter union.22,23 The 95% confidence in terval (CI) was computed for each site-specific SIR using the Byar ap proximation when the number of cases was five or more or the Fish er's exact method when the number of cases was fewer than five.24 For selected cancer sites, trends in risk by duration of carpenter union work were further investigated by calculation of directly standardized rate ratios (SRR) using procedures provided by the National Institute for Occupational Safety and Health LTAS. The LTAS calculates the SRR as a ratio of weighted rates in which the weight for each stratumspecific weight is the combined per son-years for the observed cohort across a duration of union work cat egories. In these analyses, the lowest duration of carpenter work strata (<10 years) was used as the referent group for comparison with higher duration of work strata. Taylor series approximation procedures were used to calculate 95% confidence inter vals for each SRR and tests for linear trend in the SRR with duration of carpenter union work were pre formed using procedures described by Rothman.23 Results A total of 162,520 PYARs were contributed by the 13,354 union car penters included in this study and approximately 97% of the PYARs occurred among workers less than 65 years of age. Overall SIR analyses, as well as analyses by cancer site, are given in Table 2. Union carpenters were found to have an overall cancer incidence approximately equal to the United States (SIR = 1.07; 95% CI = 0.99 -1.16); however, the dis tribution of cancers by site demon strated interesting patterns. Cancers of the digestive organs and perito neum were significantly higher than expected (SIR = 1.24; 95% CI = 1.04-1.47). Cancers of all sites of the digestive system except the peri toneum were in excess of expected; however, only rectal cancer was sig nificantly higher than expected (SIR = 1.51; 95% CI = 1.05-2.10). Significant excess risk was observed for all cancers of the respiratory sys tem (SIR = 1.52; 95% CI = 1.291.76). Cancers of the trachea, bron chus, and lung were significantly in excess (SIR = 1.45; 95% CI = 1.22-1.72) as well as cancers of other parts of the respiratory system (SIR = 4.19; 95% CI = 2.35-6.92). In these overall analyses, the only statistically significant deficit ob served was for skin melanoma (SIR = 0.43; 95% CI = 0.23- 0.72). Jicine. Unauth 1062 Cancer Incidence Among Carpenters Dement et al TABLE 2 Overall Cancer Incidence, New Jersey Carpenters 1979-2000 Cancer Site ICD-9 Codes All malignant (Mn) neoplasms Buccal cavity and pharynx Mn of lip Mn of tongue Mn of other parts of buccal cavity Mn of pharynx Digestive organs and peritoneum Mn of esophagus Mn of stomach Mn of intestine except rectum Mn of rectum Mn of biliary passages, liver, and gallbladder Mn of liver not specified as primary/secondary Mn of pancreas Mn of peritoneum and other and unspecified of digestive organs Respiratory system Mn of larynx Mn of trachea, bronchus, and lung Mn of other parts of respiratory system Male genital organs Mn of prostate Mn of testes Urinary organs Mn of kidney Mn of bladder and other urinary organs Other and unspecified sites Mn of skin melanoma Mn of eye Mn of brain and other parts of nervous system Mn of thyroid gland Mn of bone Mn of connective tissue Mn of other and unspecified sites Neoplasms of lymphatic and haematopoietic tissues Non-Hodgkins lymphoma Hodgkin's disease Leukemia and aleukemia Myeloma 140 -208.9 140 -149 140 -140.9 141-141.9 142-145.9 146 -149.9 150 -159 150 -150.9 151-151.9 152-153.9 154 -154.9 155-155.1 156 -156.9 155.2 157-157.9 158 -159.9 160 -165 161-161.9 162-162.9 160 -160.9 163-165.9 185-187 185-185.9 186 -186.9 188 -189 189.0-189.2 188.0-188.9 189.3-189.9 Mixed 172-172.9 190 -190.9 191-192.9 193-193.9 170 -170.9 171-171.9 187-187.9 194 -199.9 200 -208 200 -200.9 202-202.9 201 -201.9 204 -208.9 203-203.9 Obs 592 24 3 5 5 11 130 8 12 49 35 12 0 14 0 166 14 137 15 123 104 19 45 16 29 57 14 0 14 6 1 2 20 46 19 6 12 9 Exp 551.67 25.27 3.70 5.56 7.91 8.11 104.84 7.23 11.74 42.04 23.21 7.31 0.00 11.97 1.34 109.53 11.59 94.37 3.58 122.40 107.66 14.74 53.90 18.64 35.27 92.89 32.84 1.37 12.78 6.36 1.61 4.37 14.46 60.74 32.18 7.01 15.48 6.08 SIR 1.07 0.95 0.81 0.90 0.63 1.36 1.24 1.11 1.02 1.17 1.51 1.64 - 1.17 0.00 1.52 1.21 1.45 4.19 1.00 0.97 1.29 0.83 0.86 0.82 0.61 0.43 1.10 0.94 0.62 0.46 1.38 0.76 0.59 0.86 0.78 1.48 95% CI 0.99-1.16 0.61-1.41 0.17-2.37 0.29 -2.10 0.20-1.48 0.67-2.42 1.04 -1.47 0.48 -2.18 0.53-1.79 0.86-1.54 1.05-2.10 0.85-2.87 - 0.64-1.96 - 1.29 -1.76 0.66 -2.03 1.22-1.72 2.35-6.92 0.84-1.20 0.79-1.17 0.78 -2.01 0.61-1.12 0.49-1.39 0.55-1.18 0.46-0.80 0.23-0.72 0.60-1.84 0.34 -2.05 0.02-3.46 0.06-1.65 0.84 -2.13 0.55-1.01 0.36-0.92 0.31-1.86 0.40-1.35 0.68 -2.81 Analyses by age demonstrated that most excess cancer risk occurred among workers over 50 years of age as shown in Table 3. The SIR for all cancer was significantly reduced for workers less than 50 years of age (SIR = 0.60, P < 0.05); however, excess cancer was seen among work ers over the age of 50 (SIR = 1.18, P < 0.05). This pattern with age is consistent with the "healthy worker effort" observed in many mortality studies of occupational cohorts.22 Significant excess cancer incidence (P < 0.05) was observed for all cancers of the digestive organs and peritoneum (SIR = 1.33), rectum (SIR = 1.62), lung (SIR = 1.59), and other parts of the respiratory system (SIR = 4.88) among workers older than 50 years. Cancers of the intestine except rectum were excess among older workers (SIR = 1.22), although the excess was not statisti cally significant. To account for occupational can cer induction periods, additional analyses discounted (lagged) personyears and cancer cases prior to 15 years since initiation of work in the carpenter union. Table 4 presents results for cancer sites with signifi cant excess or deficits in the lagged iicine. Unauth JOEM Volume 45, Number 10, October 2003 TABLE 3 Cancer Incidence by Age Group for Selected Cancer Sites, New Jersey Carpenters 1979-2000 Standardized Incidence Ratios Age Group (Years) <50 >50 Overall Digestive Organs and Peritoneum 0.82 1.33* 1.24* Intestine Except Rectum 0.88 1.22 1.17 Rectum 1.00 1.62* 1.51* Trachea, Bronchus, and Lung 0.45* 1.59* 1.46* Other Respiratory System Areas 2.68 4.88* 4.21* * P < 0.05. 1063 All Cancers 0.60* 1.18* 1.07 TABLE 4 Cancer Incidence for Selected Sites 15-Year Lag from Initial Date of Carpenter Union Work, New Jersey Carpenters, 19792000 Cancer Site ICD-9 Codes Obs Exp SIR 95% CI All malignant (Mn) neoplasms Digestive organs and peritoneum Mn of intestine except rectum Mn of rectum Respiratory system Mn of trachea, bronchus, and lung Mn of other parts of respiratory system Male genital organs Mn of Testes Other and unspecified sites Mn of skin melanoma Mn of other and unspecified sites 140 -208 485 404.90 1.20 1.10 -1.31 150 -159 110 80.74 1.36 1.16 -1.60 152-153.9 42 32.52 1.29 0.98 -1.67 154 -154.9 28 17.84 1.57 1.12-2.15 160 -165 146 86.34 1.69 1.47-1.94 162-162.9 127 74.88 1.70 1.46 -1.97 160 -160.9 163-165.9 11 2.45 4.47 2.50 -7.39 185-187 100 94.37 1.06 0.89 -1.25 186 -186.9 9 3.63 2.48 1.29 -4.32 Mixed 42 45.47 0.92 0.66 -1.24 172-172.9 9 19.69 0.46 0.24-0.79 187-187.9 194 -199.9 17 10.76 1.58 1.01 -2.37 analyses. These analyses demon strated the same patterns of excess cancer risk as the overall analyses, although the SIR values were some what higher. In the lagged analyses, malignant neoplasms of the testes was significantly elevated (SIR = 2.48, 95% CI = 1.29-3.32) and mel anoma of the skin was significantly deficit (SIR = 0.46, 95% CI = 0.24 - 0.79). Duration of carpenter union work was used as a surrogate of occupa tional exposures among these work ers. Trends in the risk of cancer were analyzed by examination of the SIR values by 10-year strata of carpenter work in the 15-year lagged analyses (Table 5). Prior to 10 years of car penter union work, only cancers of other parts of the respiratory system (SIR = 4.22, P < 0.05) and testes (SIR = 2.58, P < 0.05) were signif icantly elevated. However, all other sites that were shown to be in excess in the lagged analyses demonstrated increasing trends in the SIR with duration of carpenter work, with the highest SIR values for carpenters having worked in the union for more than 30 years. In the strata greater than 30 years, the SIR for all cancers was 3.34 (P < 0.05) with very large excesses being observed for all can cers of the digestive organs and peri toneum (SIR = 3.98, P < 0.05), rectum (SIR = 4.85, P < 0.05), lung (SIR = 4.56, P < 0.05), and other parts of the respiratory system (SIR = 11.00, P < 0.05). Trends in the risk of selected can cers were further examined in the analyses using directly standardized rate ratios (SRR) and these results are shown in Table 6. Tests for linear trends in the SRR with duration of carpenter union work were signifi cant (P < 0.05) for all cancers of the digestive organs and peritoneum, lung, and other parts of the respira tory system. The SRRs for cancers of the intestine except rectum increased with work duration although the trend was not statistically significant. Although not shown in Table 6, SRRs were calculated for cancers of the rectum; however, the observed patterns of risk were not stable as a result of only one observed cancer in the reference strata (1 observed and 3.1 expected). It should be noted that all trends except cancers of other parts of the respiratory system were positive. The negative trend for can cers of other parts of the respiratory system is largely because of the sig nificant excess for cancers at these sites observed for workers with less than 10 years of union work (see Table 5). As previously discussed, cancers in this category are largely pleural cancers (including mesotheli- iicine. Unauth 1064 Cancer Incidence Among Carpenters Dement et al TABLE 5 Cancer Incidence for Selected Cancer Sites by Duration of Carpenter Union Work 15 Year Lag from Initial Date of Carpenter Union Work, New Jersey Carpenters, 1979-2000 Standardized Incidence Ratios Years of Union Work 0 -10 11-20 21-30 >30 Overall Digestive Organs and Peritoneum 0.88 1.51* 1.39 3.98* 1.36* Intestine Except Rectum 0.89 1.62* 1.06 2.73 1.29 Rectum 1.25 1.58 1.54 4.85* 1.57* Trachea, Bronchus, and Lung 1.01 1.82* 2.03* 4.56* 1.70* Other Respiratory System Areas 4.22* 1.08 10.02* 11.00* 4.47* Testes 2.58* 2.57 2.48* All Cancers 0.83* 1.15 1.40* 3.34* 1.20* * P < 0.05. TABLE 6 Directly Standardized Rate Ratios by Duration of Carpenter Union Work, New Jersey Carpenters, 1979-2000 Standardized Rate Ratio (95% Confidence Interval) Years of Union Work 0-10 11-20 21 -30 >30 Rothman Trend Test Digestive Organs and Peritoneum 1.0 2.7 (1.0-7.7) 3.5 (1.3-9.3) 3.8 (1.5-10.0) P < 0.05 Intestine Except Rectum 1.0 0.7 (0.2-3.0) 1.8 (0.6-4.9) 1.6 (0.5-4.7) P > 0.05 Trachea, Bronchus, and Lung 1.0 2.7 (0.9-7.8) 2.8 (1.0-8.1) 4.5 (1.7-12.4) P < 0.05 Other Respiratory System Areas 1.0 1.3 (0.2-9.2) 0.9 (0.2-4.0) 0.7 (0.2-2.7) P < 0.05 oma) with asbestos exposures being the most likely risk factor. Numerous studies have shown mesothelioma to be increased among workers with low cumulative exposures.25 The distribution of cancers in sev eral categories was further investi gated to determine the most promi nent sub-site. Within the category of "cancers of the intestine except rec tum" (ICD-9 codes 152-153.9), 44 of 49 cancers were in the colon (ICD-9 codes 153-153.9). In the cat egory cancers of `other parts of the respiratory system' (ICD-9 codes 160-160.9 and 163-165.9), 11 of 15 cancers were of the pleura (ICD-9 codes 163- 163.9), and 2 of 15 were of the nasal cavity (160.0). Discussion Our analyses used SEER cancer rates for the US population to calcu late expected cancers by site among the carpenter cohort. We chose to use SEER rates as opposed to rates for only New Jersey because of their general acceptance and stability of rates when stratified by age, race, sex, and calendar time. However, cancer incidence rates vary with geo graphic location, and this could have some impact on our results and in terpretations. Analyses of New Jer sey cancer incidence for 1994 through 1998 compared with SEER data for white males have shown New Jersey age-adjusted total cancer incidence rates (cases per 100,000) to be approximately 10% higher than the US (505.3 vs. 457.7). During this same time period, rates for colorectal cancer in New Jersey also were shown to be higher than the US (62.7 vs. 53.6); however, lung cancer rates where slightly lower than US rates (77.2 vs. 78.0).26 Based on these data, our analyses may have slightly overestimated the risk for all cancer sites combined and for colorectal cancers and slightly underestimated the risk of lung cancer among New Jersey carpenters. However, the dif ferences in New Jersey and SEER rates are not sufficient to explain the large and statistically significant ex cesses observed for all cancers (SIR = 3.34), digestive system can cers (SIR = 3.98), and lung cancers (SIR = 4.56) among workers in the carpenter union for more than 30 years, after controlling for a mini mum cancer latency of 15 years. These data suggest occupational ex posures as contributing factors for risks at these cancer sites. The excess risk for cancers of the trachea, bronchus and lung that we observed is consistent with findings from prior mortality studies, as well as hospital and community-based case control studies. Elevated lung cancer mortality risk estimates rang ing from 1.1 to 1.2 among carpenters and related crafts have been observed,4-7'10'n in addition to ele vated odds ratios (OR = 1.1- 1.8) when exposure to employment as a carpenter or construction worker and lung cancer have been examined.27-30 Exposures to asbestos, sil iicine. Unauth JOEM Volume 45, Number 10, October 2003 1065 ica, and wood dust, which are com mon among carpenters, have also been associated with an elevated risk for these cancers.31 An increasing trend for cancers of the respiratory system was observed, with carpen ters who were UBC members for at least 30 years to be at greatest risk. Duration of employment as a con struction worker was also explored by Stern7 who noted an elevated risk for lung cancer mortality for each 10-year period of union membership; however this trend did not reach statistical significance, which was observed in our study. Higher-thanexpected numbers of cases were noted for other respiratory cancers, which consisted mostly of cancers of the pleura. The large excess observed for pleural cancers is largely ex plained by occupational exposures to asbestos, known to occur among carpenters. 32 An increase in rectal cancer and intestinal cancer contributed to the overall increase in the incidence of cancers of the digestive tract; whereas no increased risk of stomach cancer occurred. Some studies have noted an excess in morality due to stomach cancer with risk estimates ranging from 1.2 to 1.4,6'7'11 as well as a case-control study that observed an elevated odds ratio (OR = 1.7, 95% CI = 1.0-2.7) for stomach cancer among workers exposed to wood dust for at least 15 years.33 Prior study findings with regard to cancers of the colon and rectum have been mixed, while some study find ings are difficult to use for compari son because these cancers were com bined into a single category of colorectal cancer in the analyses. El evated odds ratios of 9.4 (95% CI = 2.0- 44.7) for rectal cancer34 and 1.4 (95% CI = 0.9-2.2) for colorectal cancer33 were observed among workers exposed to wood dust. How ever, no increase in rectal cancer was observed by Dumas et al.35 who examined workers exposed to wood dust for a minimum of 15 years. Furthermore, employment as a car penter was not associated with an increased risk of either colon or rec tal cancer in a case-control study conducted by Arman et al.36 A new finding from this study includes a higher than expected inci dence of testicular cancer when a 15-year latency period was consid ered. This excess was seen only among workers with less than 10 years of union membership. Testicu lar cancer accounts for only 1% of all cancers in men, but it is the most common form of cancer in young men ages 15 to 35 years.37 Of the 19 workers with testicular cancer in the present study, 15 occurred among workers less than 45 years of age. The 5-year survival rate for testicular cancer is 93%,12 which may explain why previous mortality studies have not observed an increased risk for this cancer among carpenters or re lated trades. Pollan et al.38 observed an excess incidence of testicular can cer (nonseminoma tumors only; SIR = 1.71, 95% CI = 1.01-2.92) among construction workers in Swe den; however, these findings were not noted in other studies that used incidence data.18 The etiology of tes ticular cancer is poorly understood and only a few risk factors have been identified, including cryptorchidism (undescended testis), abnormal tes ticular development, and Klinefelter's syndrome. Occupa tional agents that have been explored as possible risk factors include plas tics, pesticides, and solvents includ ing dimethylformamide (DMF), which is commonly used in acrylic fiber spinning, chemical manufactur ing, and pharmaceutical produc- tion.39 Two small case studies ob served higher than expected cases of testicular cancer among airframe re pair shop workers,40 and finishers in a leather tannery41 who used DMF in their jobs. The degree to which DMF is used among carpenters and consid ered as a possible risk factor for this workgroup is not known. We observed a lower-than-ex- pected incidence of melanoma (SIR = 0.46) among our cohort of carpenters compared with national data. After considering a 15-year lag period, lower than expected cases were also noted when we used New Jersey State Cancer Registry data for comparison, but these findings were not statistically significant (SIR = 0.65, 95% CI = 0.30-1.23). Al though carpenters may be exposed to sunlight while at work, the pattern of exposure may reflect lower than ex pected risk estimates for melanoma. Intermittent high intensity exposure to sunlight, especially that results in sunburn has been associated with melanoma; whereas, chronic low dose exposure, which may be the more typical type of occupational exposure among carpenters, has been associated more with nonmelanoma skin cancer including basal cell and squamous cell carcinomas.42 Cigarette smoking is an estab lished cause of lung cancer and has been positively associated with can cers at numerous other sites includ ing the stomach,43 rectum,44and colon.45,46 Other lifestyle factors such as alcohol use and high dietary fat intake have also been implicated as possible risk factors for these cancers.45'47'48 Unfortunately, informa tion about these lifestyle factors among our cohort of carpenters was not available and therefore were not considered as possible confounders in the analyses. Workers in the construction indus try have been identified as heavy consumers of both tobacco and alco hol. For example, a higher preva lence of cigarette smoking was ob served among construction laborers in National Health and Nutrition Ex amination Survey III compared with the overall population of employed study participants.49 In addition, the construction industry was ranked as the third highest for alcohol use among workers ages 18 to 49 among all U.S. industry groups.50 These study findings support the impor tance of controlling for these behav iors in the analyses among carpen ters, especially when examining cancers that are associated with both lifestyle and occupational factors. iicine. Unauth 1066 Cancer Incidence Among Carpenters Dement et al However, given the significant ex cess risk of respiratory cancers and cancers of the digestive system that we observed, especially among car penters who were employed in the union for at least 30 years, it is unlikely that these elevated risk esti mates can be attributed solely to these lifestyle factors.51'52 There are numerous strengths to this study including a reasonably large cohort size and the availability of union membership and health fund participation information that allowed us to maintain follow-up of carpenters who were either active UBC members or retired and still residing in New Jersey. Another strength of the study is the linkage with a well-established cancer regis try. Furthermore, our data allowed us to examine the duration of union membership as a risk factor for se lected cancers, which has been ex amined in prior mortality studies of carpenters6'7; however, this has not been conducted in a large cohort study that used cancer incidence data. Like other studies of construc tion workers, a limitation of our study is the lack of information on specific exposures experienced by cohort members. The cohort covers most union carpenters in New Jersey but does not include nonunion car penters. The degree to which union carpenters are representative of car penters as a whole has not been investigated; however, a general ob servation is that union carpenters are more likely to be employed by larger contractors. Nationally, only 19% of carpenters are union members; how ever, this percentage is higher in New Jersey where more than 30% of construction workers are unionized.53 Conclusions Carpenters enrolled in the UBC had elevated risk estimates for can cers that have been observed in prior studies. The cancer sites with ex cesses are consistent with carpenter exposures to known and suspected carcinogens, such as asbestos, silica, and wood dust. An increased risk of rectal cancer has been suggested in prior studies; however, our study provided stronger evidence that car penters are in fact at risk, especially those who were employed for more than 30 years. Further research is needed to explore possible risk fac tors for the elevated risk of testicular cancer that we observed. Acknowledgments This work could not have been accom plished without the collaboration of the New Jersey Carpenters Funds and the valuable assistance provided by George R. Laufenberg, George Fernett, and Mary K. Willis. Toshi Abe of the New Jersey State Cancer Registry provided the cancer case data for this project and expert assistance in their interpre tation. References 1. Bureau of Labor Statistics, U. S. Depart ment of Labor. Occupational Outlook Handbook, 2002-03 Edition. Carpenters, on the Internet at http://www.bls.gov/ oco/ocos202.htm. 2. Lipscomb HJ, Dement JM, Loomis DP, Silverstein B, Kalat J. Surveillance of work-related musculoskeletal injuries among union carpenters. Am J Ind Med. 1997;32:629 - 640. 3. Blair A. Cancer rates and risks. Occupa tional Studies Section, Division of Can cer Etiology, National Cancer Institute, 2003, http://seer.cancer.gov/publications/ raterisk/risks94.html. 4. Milham S. Mortality Experience of the AFL-CIO United Brotherhood of Car penters and Joiners of America, 1969 1970. US DHHS, National Institute for Occupational Safety and Health; HEW Publication No. (NIOSH) 74-129, 1974. 5. Robinson CE, Stern F, Halperin W, et al. Assessment of mortality in the construc tion industry in the United States, 1884 1986. Am JInd Med. 1995;28:49-70. 6. Robinson CE, Peterson M, Sieber WK, Palu S, Halperin WE. Mortality of car penters' union members employed in U.S. construction or wood products in dustries, 1987-1990. Am J Ind Med. 1996;30:674-694. 7. Stern F, Schulte P, Sweeney MH, et al. Proportionate mortality among construc tion laborers. Am J Ind Med. 1995;27: 485-509. 8. Register General' s Decennial Supple ment for England and Wales, 1979 1980, 1982-83. London, UK: Occupa tional Mortality, Office of Population Censuses and Surveys, Her Majesty's Stationary Office; 1986. 9. Hrubec Z, Balir AE, Rogot E, Vaught J. Mortality risks by occupation among U.S. veterans of known smoking status, 1954 -1980. NIH Publication No. 92 3407; 1992. 10. Dement J, Lipscomb H, Epling C, Desai T, Li L, DeLarco B. Surveillance of construction workers: final report to the Center to Protect Workers' Rights CPWR/NIOSH Cooperative Agreement (CCU312014); January, 2001. 11. Milham S. Occupational mortality in Washington State, 1950-1979. US DHHS, National Institute for Occupa tional Safety and Health; Publication No. (NIOSH) 83-116; 1983. 12. Ries LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review, 1973 1999, National Cancer Institute, 2000. http://seer.cancer.gov/ csr.1973 1999.htm 13. Dement J Wood Dust. In: Bingham E, Cohrssen B, Powell CH, eds. Patty's Toxicology, Vol I. 5th ed. New York: John Wiley & Sons, Inc.; 2001:619-659. 14. Nylander LA, Dement J. Carcinogenic effects of wood dust: review and discus sion. Am J Ind Med. 1993;24:619 - 647. 15. International Agency for Research on Cancer. IARC Monographs on the Eval uation of Carcinogenic Risk to Humans: Volume 62: Wood Dust and Formalde hyde. Lyon, France: World Health Orga nization; 1995. 16. Blot WJ, Chow W, McLaughlin JK. Wood dust and nasal cancer risk: a re view of the evidence from North Amer ica. J Occup Med. 1997;39:148-156. 17. Demers PA, Kogevinal M, Boffetta P et al. Wood dust and sino-nasal cancer: pooled reanalysis of twelve case-control studies. Am JInd Med. 1995;28:151-166. 18. Hall NEL, Rosenman KD. Cancer by industry: analysis of a population-based cancer registry with an emphasis on bluecollar workers. Am J Ind Med. 1991;19: 145-159. 19. Ferlay J. IARCtools Version 2.01, Lyon, France: Unit of Descriptive Epidemiol ogy, International Agency for Research on Cancer, World Health Organization; 2001. 20. Steenland K, Beaumont J, Spaeth S, et al. New developments in the life table anal ysis system of the National Institute for Occupational Safety and Health. J Occup Med. 1990;32:1091-1098. 21. Cassinelli R, Kock KJ, Steenland K, Spaeth S, Laber P. User Documentation--PC LTAS: Life Table Analysis System for Use on the PC. Cincinnati, Jicine. Unauth JOEM Volume 45, Number 10, October 2003 1067 OH: USDHHS, Centers for Disease Con trol and Prevention, National Institute for Occupational Safety and Health; 2001. 22. Checkoway H, Pearce N, CrawfordBrown DJ. Research Methods in Occu pational Epidemiology. New York: Ox ford University Press; 1989:153-155. 23. Rothman KJ, Greenland, S. Modern Ep idemiology. 2nd ed. Philadelphia: Lippincott, Williams & Wilkins; 1998:15. 24. Rothman KJ, Boice JD. Epidemiologic Analysis With A Programmable Calcula tor. Washington, DC: US Government Printing Office; 1979. 25. Rodelsperger K, Jockel KH, Pohlabeln H, Romer W, Woitowitz HJ. Asbestos and man-made vitreous fibers as risk factors for diffuse malignant mesotheli oma: results from a German hospitalbased case-control study. Am J Ind Med. 2001;36:262-275. 26. Burger SS, Klotz JB, Weinstein R, Abe T, Van Loon S. Cancer incidence in New Jersey 1995-1999: implementation of the year 2000 population standard. New Jer sey Department of Health and Senior Services. Cancer Epidemiology Services. New Jersey State Cancer Registry; Sep tember, 2001. 27. Burns, PB, Swanson, GM. The occupa tional cancer incidence surveillance study: risk of lung cancer by usual occu pation and industry in the Detroit metro politan area. Am J Ind Med. 1991;19: 655-671. 28. Coggon D, Pannett B, Osmond C, Acheson ED. A survey of cancer and occupation in young and middle aged men I: cancers of the respiratory tract. Br J Ind Med 1986;43:332-338. 29. Hoar-Zahm S, Brownson R, Chang J, Davis J. Study of lung cancer histologic types, occupation, and smoking in Mis souri. Am J Ind Med. 1989;15:565-578. 30. Jockel KH, Ahrens W, Wichmann HE, et al. Occupational and environmental haz ards associated with lung cancer. Int J Epidemiol. 1992;21:202-213. 31. Blot WJ, Davies JE, Brown LM, et al. Occupation and the high risk of lung cancer in Northeast Florida. Cancer. 1982;50:365-371. 32. Lipscomb HJ, Dement JM. Respiratory diseases among union carpenters: cohort and case-control analyses. Am J Ind Med. 1998;33:131-150. 33. Siemiatycki J, Richardson L, Gerin M, et al. Associations between several sites of cancer and nine organic dusts: results from an hypothesis-generating casecontrol study in Montreal, 1979 -1983. Am J Epidemiol 1986;123:235-249. 34. Peters RK, Garabrant DH, Yu MC, Mack TM. A case-control study of occupational and dietary factors in colorectal cancer in young men by subsite. Cancer Res. 1989; 49:5459-5468. 35. Dumas S, Parent, ME, Siemiatycki J, Brisson J. Rectal cancer and occupational risk factors: a hypothesis-generating, ex posure-based case-control study. Int J Cancer. 2000;87:874-879. 36. Arbman G, Axelson O, Fredriksson M, Nilsson E, Sjodahs R. Do occupational factors influence the risk of colon and rectal cancer in different ways? Cancer. 1993;72:2543-2549. 37. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1988. Cancer J Clin. 1998;48:6. 38. Pollan M, Gustavsson P, Cano MI. Inci dence of testicular cancer and occupation among Swedish men gainfully employed in 1970. Ann Epidemiol. 2001;118:554562. 39. NIOSH: Preventing adverse health ef fects from exposure to Dimethylformamide (DMF). USDHHS, Centers for Disease Control, National Institute for Occupational Safety and Health. NIOSH Publication No.1990:90-105. 40. Ducatman AM, Conwill DE, Crawl J. Germ cell tumors of the testicle among aircraft repairmen. J Urol. 1986;136: 834 - 836. 41. Levin SM, Landrigan PJ, Baker DB, Monaghan SV. Testicular cancer in leather tanners exposed to dimethylformamide. Lancet. 1987;2:1154. 42. Armstrong BK, Kricker A. The epidemi ology of UV induced skin cancer. Potochem Photobiol. 2001;63:8-18. 43. Chao A, Thun MJ, Henley SJ, Jacobs EJ, McCullough ML, Calle EE. Cigarette smoking, use of other tobacco products and stomach cancer mortality in U. S. adults: the cancer prevention study II. Int J Cancer. 2002;1014:380-389. 44. Heineman EF, Hoar Zahm S, McLaugh lin JK, Vaught JB. Increased risk of colorectal cancer among smokers: results of a 26-year follow-up of U.S. veterans and a review. Int J Cancer. 1995;59:728 738. 45. Le Marchand L, Wilkens LR, Kolonel LN, Hankin JH, Lyu LC. Associations of sedentary lifestyle, obesity, smoking, al cohol use, and diabetes and the risk of colorectal cancer. Cancer Res. 1997;57: 4787-4794. 46. Slattery ML, Potter JD, Friedman GD, Ma KN, Edwards S. Tobacco use and colon cancer. IntJ Cancer. 1997;70:259 264. 47. Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Ascherio A, Willett WC. Intake of fat, meat, and fiber in relation to risk of colon cancer in men. Cancer Res. 1994;54:2390 -2397. 48. Wynder EL, Hebert JR, Kabat GC. As sociation of dietary fat and lung cancer. JNatl Cancer Inst. 1987;794:631-637. 49. Bang KM, Kim JH. Prevalence of ciga rette smoking by occupation and industry in the United States. Am J Ind Med. 2001;40:233-239. 50. U.S. Department of Commerce. Drug use among U.S. workers: prevalence and trends by occupation and industry cate gories. USDHHS. Substance Abuse and Mental Health Services. DHHS Publica tion No. (SMA) 96-3089; 1996. 51. Axelson O. Aspects of confounding in occupational health epidemiology. Scand J Work Environ Health. 1978;4:85-89. 52. Flanders WD, Khoury MJ. Indirect as sessment of confounding: graphic de scription and limits on effect of adjusting for covariates. Epidemiology. 1990;1: 239-246. 53. Sullivan EC, Englund A, Dunlop JT, Frankowski R. The Construction Chart Book: The U.S. Construction Industry and Its Workers, Third Edition, Center to Protect Workers' Rights. Silver Spring, MD: September 2002. Jicine. Unauth