The effect of disease misclassification on the ability to detect a gene-environment interaction: implications of the specificity of case definitions for research on Gulf War illness

Table 1 Strength of the association of the PON1 Q192R-nerve agent GxE interaction with the various GWI case definitions

GWI case definition	Number of veterans classified as GWI in case–control sample^a	aRERI for GxE interaction (95% CI)^b
Research	508	7.69 (2.71 – 19.13)
Research variant 1	94	3.66 (0.17 – 13.66)
Research variant 2	206	15.84 (5.36 – 56.31)
Research variant 3	163	7.93 (2.38 – 24.16)
CDC	1109	2.92 (0.96 – 6.38)
CDC, not Research^c	624	1.98 (0.42 – 4.67)
CDC mild-to-moderate	757	2.41 (0.68 – 5.46)
CDC mild-to-moderate, not Research^c	521	2.08 (0.46 – 4.92)
CDC Severe	352	7.16 (2.34 – 19.18)
CDC Severe, not Research^c	103	2.11 (-0.85 – 8.95)
Modified Kansas without exclusions	1090	3.84 (1.30 – 8.52)
Modified Kansas without exclusions, not Research^c	602	2.79 (0.68 – 6.67)
Modified Kansas with exclusions	748	3.42 (1.20 – 7.56)
Modified Kansas with exclusions, not Research^c	509	2.92 (0.85 – 6.82)

Abbreviation: aRERI adjusted relative excess risk due to interaction, CI confidence interval, GWI Gulf War illness, PON1 Q192R polymorphism of the paraoxonase-1 gene
^aThe control group included 508 deployed veterans who satisfied none of these case definitions
^bThe potential confounders controlled for in the adjusted models included: age, sex, service branch, rank, active duty vs Guard/Reserve, special strata, and combat exposure scale. Detailed tables showing the calculation of the aRERI are given in Tables S8-S21
^cThese groups contain those personnel meeting the CDC or Modified Kansas case definitions but excluding those also meeting the Research definition

ISSN: 1471-2288