Table 4.

Association of PFOS with total cholesterol is dependent on extent of lipid lowering

Non-responders Covariate	Estimated regression coefficient	Standard Error	p-value
Intercept	196.0	27.2	<.001
PFOSpost (log)	−0.226	3.20	0.944
Sex	17.4	6.21	0.006
Age	0.337	0.200	0.094
Education level	−0.278	0.958	0.772
Race	−1.39	4.63	0.764
BMIpost	−0.326	0.359	0.366
Smoking history	−4.83	3.43	0.160
Years living in current county	0.278	0.147	0.061
Responders Covariate	Estimated regression coefficient	Standard Error	p-value
Intercept	151	32.8	<.001
PFOSpost (log)	9.57	3.02	0.002
Sex	14.2	6.14	0.022
Age	0.163	0.223	0.467
Education level	0.243	0.995	0.807
Race	−5.38	5.94	0.367
BMIpost	0.009	0.405	0.983
Smoking history	−1.38	3.07	0.653
Years living in current county	0.009	0.157	0.954

^*Responders lost >6 mg/dL total cholesterol (median loss). (n=171 non responders, n=162 responders).

BMI: Body mass index. PFOS= Perfluorooctanesulfonic acid

The estimates above define a formula for the prediction of total cholesterol in subject’s post intervention.

For variable “sex”, 0=male. 1=female; For variable “Race”, 1=African American. 2=Caucasian. 3=Asian. 4=Hispanic/Latino. 5=American Indian/Alaskan Native. 6= Native Hawaiian/other Pacific. For variable “Smoking history”, 1=current smoker. 2=Stopped within 1 year. 3=stopped more than 1 year ago. 4=never smoked.