Table 2.
RR estimates for kidney cancer associated with TCE exposure (overall and highest exposure group) from cohort and case-control studies.
| Overall Exposure | Highest TCE Exposure Group | |||||
|---|---|---|---|---|---|---|
| Study | RR (95% CI) | Alternate RR estimates | Exposure Category | RR (95% CI) | Alternate RR estimates | Comments |
| Cohort Studies | ||||||
| Anttila et al. [38] | 0.87 (0.32, 1.89) | None | 100+ μmol/L U-TCA a | 1.0 assumed b | ICD-7 180. SIR. Reported high exposure group results for some cancer sites but not kidney. | |
| Axelson et al. [37] | 1.16 (0.42, 2.52) | 1.07 (0.39, 2.33) with estimated female contribution to SIR addedc | ≥2 year exposure and 100+ mg/L U-TCA | 1.0 assumed b | ICD-7 180. SIR reported for males only, but there was a small female component to the cohort. Reported high exposure group results for some cancer sites but not kidney. | |
| Boice et al. [41] | 0.99 (0.4, 2.04) | None | ≥5 years exposure | 0.69 (0.22, 2.12) | None | ICD-9 189.0−189.2. Overall exposure SMR for potential routine exposure; results for any potential exposure not reported. Mortality RR for highest exposure group for potential routine or intermittent exposure, adjusted for date of birth, dates 1st and last employed, race, and sex; referent group is workers not exposed to any solvent. |
| Greenland et al. [42] | 0.99 (0.30, 3.32) | None | NA | b | ICD-8 codes not specified, presumably all of 189. Mortality OR from nested case-control study. | |
| Hansen et al. [39] | 1.1 (0.3, 2.8) c | None | ≥1,080 months × mg/m3 | 1.0 assumed b | ICD-7 180. SIR. Reported high exposure group results for some cancer sites but not kidney. | |
| Morgan et al. [43] | 1.14 (0.51, 2.58) (Morgan et al. [57]) | 1.32 (0.57, 2.6) Published SMR | High cumulative exposure score | 1.59 (0.68, 3.71) | 1.89 (0.85, 4.23) for medium/high peak | ICD-7 180, ICD-8, -9 189.0−189.2. Overall mortality RR from Morgan et al. [57]. RRs adjusted for age and sex. |
| Raaschou-Nielsen et al. [40] | 1.20 (0.94, 1.50) | 1.20 (0.98, 1.46) for ICD-7 180; c 1.4 (1.0, 1.8) for subcohort with expected higher exposures |
≥5 years in subcohort with expected higher exposure levels | 1.7 (1.1, 2.4) | 1.6 (1.1, 2.2) for ≥5 years in total cohort; c 1.4 (0.99, 1.9) ICD-7 180 ≥5 years in total cohort c |
ICD-7 180.0 (RCC). |
| Radican et al. [27] | 1.18 (0.47, 2.94) | None | >25 unit-years | 1.11 (0.35, 3.49) d | Incidence RR: 0.9 (0.3, 3.2) (Blair et al. [44]) d | ICD-8, -9 189.0, ICD-10 C64. Mortality RR adjusted for age, sex and race, with workers with no chemical exposures as referent group. |
| Zhao et al. [25] | 1.7 (0.38, 7.9)e | Incidence RR: 2.0 (0.47, 8.2); e Mortality RR no lag: 0.89 (0.22, 3.6); e Incidence RR no lag : 2.1 (0.56, 8.1);e SMR: 2.22 (0.89, 4.57) (Boice et al. [26]) |
High exposure score | 7.40 (0.47, 116) | Mortality RR: 1.82 (0.09, 38.6); Incidence RR no lag: 7.71 (0.65, 91.4); Mortality RR no lag: 0.96 (0.09, 9.91); Mortality RR: 2.12 (0.63, 7.11) for ≥5 years as test stand mechanic (Boice et al. [26]); 3.13 (0.74, 13.2) for ≥4 test-year engine flush (Boice et al. [26]) |
ICD-9 189. Mortality RR for males only for overall exposure with 20-year lag; adjusted for age, SES, time since first employment, exposure to other carcinogens. Overall mortality results reflect same number exposed cases (10 with no lag) as do incidence results. Overall RRs estimated by combining across exposure groups. Incidence RR for highest TCE exposure group reflects more exposed cases than does the mortality results and is used in primary analysis. Boice et al. [26] cohort, with seven exposed deaths, overlaps Zhao et al. [25] cohort. |
| Case-Control Studies: f | ||||||
| Brüning et al. [58] | 2.47 (1.36, 4.49) | 1.80 (1.01, 3.20) for longest job held in industry with TCE exposure | ≥20 years self-assessed exposure | 2.69 (0.84, 8.66) | None | RCC. OR for self-assessed TCE exposure adjusted for age, sex, and smoking. |
| Charbotel et al. [50] | 1.88 (0.89, 3.98) | 1.64 (0.95, 2.84) for full study; 1.68 (0.97, 2.91) for full study with 10-year lag | High cumulative dose | 3.34 (1.27, 8.74) | 3.80 (1.27, 11.40) for high + peaks; Full study, high: 2.16 (1.02, 4.60) + peaks: 2.73 (1.06, 7.07); Full study with 10-year lag, high: 2.16 (1.01, 4.65) + peaks: 3.15 (1.19, 8.38); Full study, additional adjustment, high: 1.96 (0.71, 5.37) + peaks: 2.63 (0.79, 8.83) |
RCC. ORs for subgroups with good confidence about exposure assessment. Matched on sex and age, and adjusted for smoking and BMI. Highest exposure group alternate estimates with additional adjustment were also adjusted for exposure to cutting fluids and other petroleum oils. |
| Dosemeci et al. [51] | 1.30 (0.9, 1.9) | None | NA | b | RCC. OR adjusted for age, sex, smoking, hypertension and/or use of diuretics and/or anti-hypertension drugs, BMI. | |
| Moore et al. [24] | 2.05 (1.13, 3.73) | 1.63 (1.04, 2.54) for all subjects | ≥1.58 ppm × years | 2.23 (1.07, 4.64) | 2.02 (1.14, 3.59) for all subjects | RCC. Subgroup with high-confidence assessments. OR adjusted for age, sex, and center. |
| Pesch et al. [52] | 1.24 (1.03, 1.49) | 1.13 (0.98, 1.30) with German JEM | Substantial | 1.4 (0.9, 2.1) d | 1.2 (0.9, 1.7) for JEM d | RCC. JTEM approach. Crude ORs and CIs for overall TCE exposure calculated from data provided by Pesch [59], as described in methods section. ORs for highest exposure group adjusted for age, study center, and smoking. |
| Siemiatycki [53] | 0.8 (0.3, 2.2) | None | Substantial | 0.8 (0.2, 3.4) | None | “Kidney cancer.” SE and 95% CI calculated from reported 90% CI. OR for males only, adjusted for age, income, and cigarette smoking index. |
Mean personal trichloroacetic acid in urine. 1 μmol/L = 0.1634 mg/L.
Anttila et al. [38], Axelson et al. [37] and Hansen et al. [39] report a RR estimate for highest TCE exposure groups and other cancers, but not kidney. A risk estimate of 1.0 is assigned for highest exposure in these studies to account for potential publication bias. For the SE (of the log RR) estimates for these null values, SE estimates from cancer types in the highest exposure group that were expected to have similar numbers of cases were generally used (See Appendix C of U.S. EPA [60] for further details). For Greenland et al. [42] and Dosemeci et al. [51], a risk of 1.0 is not assumed for highest exposure since only overall results are presented in those studies.
Male and female results combined assuming Poisson distribution. Details of the approach used to estimate the female contribution for Axelson et al. [37] are presented in U.S. EPA [60].
Male and female results combined using inverse-variance weighting, as in a fixed-effect meta-analysis.
To derive an overall RR estimate, results were combined across exposure groups using inverse-variance weighting, under assumptions of group independence, although the exposure groups share a referent group and, hence, are not actually independent.
The RR estimates are all ORs for incident cases.
BMI = body mass index, CI = confidence interval, cum = cumulative, ICD = International Classification of Diseases, JTEM = job-task-exposure matrix, NA = not available, OR = odds ratio, RCC = renal cell carcinoma, RR = relative risk, SE = standard error, SES = socioeconomic status, SIR = standardized incidence ratio, SMR = standardized mortality ratio, TCE = trichloroethylene, U-TCA = urinary trichloroacetic acid.