Is the cholesterol-perfluoroalkyl substance association confounded by dietary fiber intake?: a Bayesian analysis of NHANES data with adjustment for measurement error in fiber intake

Table 1 Coefficients^a (error estimates^b) for fiber(ln[g/d])^c and PFOA (ng/ml) in multivariable models^d of ln(serum cholesterol), according to fiber variable included in model and type of model; n = 7,242

Fiber variable^c included in model	ln-transformed energy intake-adjusted fiber (ln[g/d])		PFOA (ng/ml)
	OLS	Bayesian	OLS	Bayesian
None	N/A	N/A	2.59 (0.75)	2.66 (0.71)
Total fiber	-1.92 (1.58)	-2.35 (1.58)	2.53 (0.75)	2.59 (0.72)
Soluble fiber	-2.90 (1.06)	-3.53 (1.49)	2.45 (0.75)	2.49 (0.72)

^aTo simplify the presentation coefficients and error estimates were multiplied by 10^–3 for PFOA and 10^–2 for fiber
^bOrdinary least squares (OLS) error estimates are standard errors and Bayesian error estimates are standard deviations
^cModeled as ln-transformed energy intake-adjusted fiber
^dAll models were adjusted for the secondary covariates energy intake, saturated fat intake, dietary cholesterol intake, age, sex, race/ethnicity (5 categories), income to poverty ratio, smoking (4 categories, treated as an ordinal variable), and a linear and a quadratic wave variable

ISSN: 1476-069X