Potential pitfall in using cumulative exposure in exposure-response relationships: demonstration and discussion

Area of Science:

• Epidemiology
• Occupational Health
• Biostatistics

Background:

• Cumulative exposure is a common metric in quantitative epidemiologic studies.
• A known issue is the imposed symmetry between exposure duration and intensity.
• Less recognized is that a found exposure-response relationship may not reflect accurate exposure estimation.

Purpose of the Study:

• To investigate the validity of cumulative exposure as a metric in epidemiologic studies.
• To assess whether a significant exposure-response relationship implies accurate exposure estimation.
• To evaluate the reliability of quantitative risk assessment using cumulative exposure.

Main Methods:

• A simulation study based on a nested case-control analysis of mesothelioma in asbestos cement workers.
• Random assignment of exposure intensity (0.1-40 fibers/ml) to subjects.
• Logistic regression analysis to assess associations between mesothelioma risk, exposure intensity, and cumulative exposure.

Main Results:

• No association was found between mesothelioma risk and randomly assigned exposure intensity.
• Despite random intensity, 86% of trials showed a significant association between mesothelioma risk and cumulative exposure.
• This indicates a potential for misleading exposure-response relationships.

Conclusions:

• Finding a strong exposure-response relationship using cumulative exposure does not guarantee accurate exposure estimation.
• Quantitative risk assessment based solely on cumulative exposure may be unreliable.
• Confidence in risk assessment increases with multiple independent studies.