The lack of causality between asbestos and colorectal cancer

Our plain-English summary of The lack of causality between asbestos and colorectal cancer, by W. Weiss (Journal of Occupational and Environmental Medicine, 1995).

The study

• Author: W. Weiss
• Journal: Journal of Occupational and Environmental Medicine
• Published: 1995;37(12):1364-73
• Type of article: Meta-analysis
• Cited by: 11 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

To evaluate a potential causal relationship between asbestos exposure and the risk of colorectal cancer, 30 cohort studies published in English were reviewed and analyzed. The latest report of each cohort up through 1993 was used. The outcome measure was the standardized morbidity (incidence) or mortality ratio (SMR). Summary SMRs were examined by type of cohort registration, type of work, type of asbestos, latency from onset of exposure to period of risk, and incidence only. The overall relative risk was 0.99 for all 30 studies. The evidence does not meet the important criteria for a judgment of causality because the relative risk is not consistently elevated, weak in the two studies with a statistically significant elevation of SMR, and limited data do not support a dose-response relationship.

In plain English

In this review, the author pooled 30 English-language cohort studies of asbestos-exposed workers, published through 1993, to judge whether asbestos causes colorectal cancer. The cohorts, totalling 126,395 subjects, included miners, factory, shipyard and insulation workers. The author compared summed observed and expected cases or deaths (O/E). Across all 30 cohorts the overall relative risk was .99 (95% confidence interval .92 to 1.07). It was .96 (.88 to 1.05) in 16 inception cohorts, 1.07 (.94 to 1.22) in 14 cross-sectional cohorts and 1.01 (.89 to 1.15) in eight incidence studies, none statistically significant. Five chrysotile-only cohorts had a significant O/E of .87 (.77 to 0.99). The single amosite-only cohort, of factory workers, had a significant 1.85, with its interval printed as 1.16 to 2.79 in one table and 1.19 to 2.85 in another. Twelve mixed-exposure cohorts had a significant 1.15 (1.02 to 1.30), and the single insulator cohort a significant 1.59 (1.21 to 2.91). The author concluded that causality criteria were not met: the association was very inconsistent, weak when present, and lacked good dose-response evidence. The author noted missing exposure data, publication bias and little control for confounders.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Review and pooled analysis of 30 published cohort studies, using summary ratios of observed to expected cases or deaths (SMR, O/E)
• Who was studied: 30 cohorts of asbestos-exposed workers totalling 126,395 subjects, using the latest report of each through 1993: 16 inception cohorts (71,111 subjects) and 14 cross-sectional cohorts in 13 reports (55,284 subjects)
• Exposure: Occupational asbestos exposure among miners and millers, factory workers, shipyard workers, railroad shop workers, marine workers, insulators, welders, seamen and others. Fiber types were chrysotile, amosite, crocidolite, mixed types, or not described.
• Main results:
  • The overall relative risk of colorectal cancer for all 30 cohorts was .99, with a 95% confidence interval of .92 to 1.07.
  • The summary O/E for the 16 inception cohorts was .96 (95% confidence interval .88 to 1.05), not statistically significant.
  • The summary O/E for the 14 cross-sectional cohorts was 1.07 (95% confidence interval .94 to 1.22), not statistically significant.
  • The summary O/E for the eight incidence studies was 1.01 (95% confidence interval .89 to 1.15), not significant.
  • Five cohorts with only chrysotile exposure had an O/E of .87 (95% confidence interval .77 to 0.99), which was significant.
  • The single study of factory workers with only amosite exposure (Seidman et al) showed a significant SMR of 1.85. Table 6 gives its 95% confidence interval as 1.16 to 2.79, while Table 2 gives 1.19 to 2.85.
  • Twelve studies of workers exposed to mixed types of asbestos had a slightly elevated, significant SMR of 1.15 (95% confidence interval 1.02 to 1.30). The 11 studies that did not specify asbestos type had an SMR of .91, not significant.
  • The single insulator cohort (Selikoff et al) had a significant elevation of the O/E to 1.59 (95% confidence interval 1.21 to 2.91).
  • Among cohorts with data for 20 or more years after onset of exposure, the six inception cohorts had a significantly low summary O/E of .85. The four cross-sectional cohorts had a significantly elevated 1.29, dominated by the American-Canadian insulator cohort. The summary O/E for all 10 was .93, not significant.
  • In Quebec chrysotile miners and millers, O/Es for colorectal cancer by cumulative dust exposure were .58 (low), 0.71 (medium), 1.00 (high) and .99 (very high). The original authors called these patterns difficult to interpret.
• The authors' conclusion: The author concluded that the important criteria for a causal relationship between asbestos and colorectal cancer were not satisfied. The association was very inconsistent and weak when present, and there was no good evidence for a dose-response relationship.
• Limitations the authors state:
  • The author stated that most reports lacked quantitative exposure data.
  • The author stated that publication bias was a serious problem.
  • The author stated that the cohort studies made little or no effort to control for confounders such as family history and diet.
  • The author stated that cross-sectional registration of cohorts gives more opportunity for bias.
  • The author stated that the summary method depends on the validity of the component studies and could obscure useful differences between them.

Topics

The National Library of Medicine indexes this study under subjects our site covers: Colon cancer.

Medical terms in this study, explained in our Lexicon: Colorectal Cancer, Asbestos Exposure, Latency, Relative Risk, Dose-response relationship.

Related studies

• Asbestos and colorectal cancer — W. Weiss (1990)
• Evidence for excess colorectal cancer incidence among asbestos-exposed men in the Beta-Carotene and Retinol Efficacy Trial — O.A. Aliyu, M.R. Cullen, et al. (2005)
• Colorectal cancer in asbestos cement workers in Denmark — E. Raffn, E. Villadsen, et al. (1996)
• Colorectal cancer and asbestos exposure-an overview — Q. Huang, Y.J. Lan (2020)
• Occupational exposures and colorectal cancers: a quantitative overview of epidemiological evidence — E. Oddone, C. Modonesi, et al. (2014)
• Exposure to asbestos and the risk of colorectal cancer mortality: a systematic review and meta-analysis — K. Kwak, D. Paek, et al. (2019)

Browse more research in Journal of Occupational and Environmental Medicine, or return to medical research.