Comment on 'The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005)': the effect of left censoring

Our plain-English summary of Comment on 'The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005)': the effect of left censoring, by A. Farioli, S. Mattioli, S. Curti, et al (British Journal of Cancer, 2014).

The study

• Authors: A. Farioli, S. Mattioli, S. Curti, F.S. Violante
• Journal: British Journal of Cancer
• Published: 2014;111(11):2197-8
• Type of article: Comment
• Cited by: 5 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Summary of the article

This letter to the editor responded to Frost's 2013 study of mesothelioma latency among a cohort of British asbestos workers, which used a follow-up window of 1978–2005. Frost had reported a median latency of 48 years for deaths among workers first exposed before 1940 and 11 years for those first exposed from 1980 onwards. The authors argued that this pattern could reflect left censoring, because deaths before 1978 were excluded, and right censoring, because deaths after 2005 were not observed. To show the possible effect, they ran a simulation. It assumed a true latency that was the same for every period of first exposure, following a gamma distribution with a median of 32 years. The simulation used six first-exposure categories with 300 mesothelioma deaths each. With complete follow-up, time ratios ranged from 0.97 to 1.02. With left and right censoring applied, median latency fell from 47.8 to 18.8 years and time ratios fell from 1.00 to 0.37, a spurious association. Results from 10,000 repetitions were reported as highly consistent. The authors suggested that this bias could also affect Frost's estimates for duration of exposure and occupation. They asked for a reanalysis restricted to cases first exposed in 1950–1969, using the full 1972–2005 follow-up and without adjusting for year of first exposure.

PubMed has no abstract for this study, and the article doesn't print one. We wrote this summary from the full article with the help of AI. It is not the authors' wording.

In plain English

This was a letter to the editor, not a new study of workers. The authors commented on a 2013 study by Frost of mesothelioma latency (the time from first asbestos exposure to mesothelioma death) among British asbestos workers followed from 1978 to 2005. That study had reported a median latency of 48 years for deaths among workers first exposed before 1940, and 11 years for cases first exposed from 1980 on. The authors argued that the limited follow-up window could create this pattern. Deaths before 1978 were left out (left censoring), and deaths after 2005 could not be observed (right censoring). To show this, they ran a computer simulation with 300 mesothelioma deaths in each of six groups by year of first exposure. Every group was given the same true latency, with a median of 32 years. With complete follow-up, the groups' median latencies were similar, 31.2 to 32.7 years. With both kinds of censoring applied, the median fell from 47.8 years in the earliest group to 18.8 years in the latest, with time ratios falling from 1.00 to 0.37. The authors concluded that a strong bias might affect Frost's findings, and they proposed a reanalysis.

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: Letter to the editor commenting on a published study, containing a computer simulation study
• Who was studied: Simulated data only: six categories of year of first asbestos exposure, matching Frost's article, with 300 simulated mesothelioma deaths assumed in each. The mid-point years 1935, 1945, 1955, 1965, 1975 and 1985 were used as the first year of exposure. The letter comments on Frost's cohort of British asbestos workers followed from 1978 to 2005.
• Exposure: Year of first occupational exposure to asbestos (no fiber types specified). The true latency was assumed to be the same for every category, following a gamma distribution (shape 11, scale 3) with a median of 32 years.
• Main results:
  • Frost's original study (not this letter) reported a median latency of 48 years for deaths among workers first exposed before 1940, and 11 years for cases first exposed not earlier than 1980.
  • In the simulation with complete follow-up and no censored cases, the categories did not differ. Median latencies ranged from 31.2 to 32.7 years and time ratios from 0.97 to 1.02 (reference 1.00).
  • With left and right censoring applied, a spurious strong association appeared between year of first exposure and time to mesothelioma death. In the earliest category, 49 deaths were observed and the median latency was 47.8 years (time ratio 1.00, reference). In the latest category, 38 deaths were observed and the median latency was 18.8 years (time ratio 0.37).
  • The figure shows a single simulation. The authors reported that a further analysis based on 10 000 repetitions gave highly consistent findings, but those data were not shown.
• The authors' conclusion: The authors concluded that their simulated estimates were not far from those in Table 2 of Frost's article, which suggested that a strong bias might affect Frost's findings. They also suggested that this bias could distort the multivariate time ratios in Frost's Table 3. They invited Frost to reanalyse the data: include only cases first exposed between 1950 and 1969, use the entire 1972 to 2005 follow-up, and avoid adjusting for year of first exposure.

Topics

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

Medical terms in this study, explained in our Lexicon: Latency, Mesothelioma.

Related studies

• Comment on 'The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005)': methodological problems with case-only survival analysis — D. Consonni, F. Barone-Adesi, et al. (2014)
• Comment on 'The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005)' — D. Mirabelli, D. Zugna (2014)
• Response to comment on 'The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005)' — G. Frost (2014)
• The latency period of mesothelioma among a cohort of British asbestos workers (1978-2005) — G. Frost (2013)
• Occupational exposures to cosmetic talc and risk of mesothelioma: an updated pooled cohort and statistical power analysis with consideration of latency period — G.M. Marsh, A.M. Ierardi, et al. (2019)
• Letter concerning: Occupational exposures to cosmetic talc and risk of mesothelioma: an updated pooled cohort and statistical power analysis with consideration of latency period by Gary M. Marsh et al. (Inhal Toxicol. 2019 Aug 5:1-11. doi:10.1080/08958378.2019.1645768) — M.M. Finkelstein (2019)

Browse more research in British Journal of Cancer, or return to medical research.