Sample size requirements in case-only designs to detect gene-environment interaction

Area of Science:

• Epidemiology
• Genetic Epidemiology
• Biostatistics

Background:

• Molecular genetic technologies are advancing, increasing the potential for studies on gene-environment interactions in disease etiology.
• Gene-environment interactions are crucial for understanding complex diseases.
• The case-only design is a proposed efficient method for screening these interactions.

Purpose of the Study:

• To derive a method for estimating sample size requirements for case-only studies investigating gene-environment interaction.
• To compare the efficiency of case-only designs versus case-control designs for detecting gene-environment interaction.
• To present sample size estimates for gene-environment interaction detection.

Main Methods:

• Development of a sample size calculation method for case-only studies.
• Comparison of minimum sample size requirements between case-only and case-control designs.
• Illustration of sample size estimation using available population data on marginal effects.

Main Results:

• The case-only design is demonstrated to be more efficient than the case-control design for detecting gene-environment interaction, assuming independence between exposure and genotype.
• Sample size estimation methods are presented for various scenarios.
• A method is shown for sample size estimation when marginal effects of exposure and genotype are known.

Conclusions:

• The case-only design is a statistically valid and efficient approach for screening gene-environment interactions.
• This design can reduce sample size requirements compared to case-control studies under certain assumptions.
• The derived methods facilitate planning future genetic epidemiology studies focused on gene-environment interactions.