Perfluorooctanoate and changes in anthropometric parameters with age in young girls in the Greater Cincinnati and San Francisco Bay Area

Abstract

Methods: We conducted a study of per- and polyfluoroalkyl substance biomarkers, including PFOA, in girls from Greater Cincinnati (CIN, N = 353) and the San Francisco Bay Area (SFBA, N = 351). PFOA was measured in the baseline serum sample collected in 2004-2007 of 704 girls at age 6-8 years. Mixed effects models were used to derive the effect of PFOA on BMI, waist-to-height and waist-to-hip ratios over increasing age in this longitudinal cohort.

Results: Median PFOA serum concentrations were 7.3 (CIN) and 5.8 (SFBA) ng/mL, above the U.S. population median for children 12-19 years in 2005-2006 (3.8 ng/mL). Log-transformed serum PFOA had a strong inverse association with BMIz in the CIN girls (p = 0.0002) and the combined two-site data (p = 0.0008); the joint inverse effect of PFOA and Age*PFOA weakened at age at 10-11 years. However, in the SFBA group alone, the relationship was not significant (p = 0.1641) with no evidence of changing effect with age. The effect of PFOA on waist:height ratio was similar to BMIz at both sites, but we did not find a significant effect of PFOA on waist:hip ratio in either the CIN or SFBA girls.

Conclusions: PFOA is associated with decreased BMI and waist:height ratio in young girls, but the strength of the relationship decreases with age. Site heterogeneity may be due to greater early life exposure in Cincinnati.

Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. Use of trade names is for identification only and does not imply endorsement by the CDC, the Public Health Service, or the US Department of Health and Human Services.

Figure 1:. Effect of PFOA expoexposure on BMI¹ (quintile β estimates)²

¹The effects of PFOA and other covariates on BMIz were derived from a model which contained race/ethnicity, parent education, PFOA, age at examination, average kilocalories in prior 12 months, hours of moderate or vigorous physical activity in last 12 months ² PFOA exposure was modeled as quintile groups of the values of the PFOA biomarker CIN=Cincinnati; SFBA-San Francisco Bay Area

Figure 2:. Effect estimates (β values) of the combined effect of PFOA and Age on BMIz

¹ The effects of PFOA and other covariates on BMIz were derived in a model which contained race/ethnicity, parent education, PFOA, age at examination, average kilocalories in prior 12 months, hours of moderate or vigorous physical activity in last 12 months ²Age was modeled in four categories (6–7, 8–9, 10–11, 12–13 and 14 or older) with age 8–9 serving as the comparison group. The combined effect of PFOA and age for each age group was calculated as βPFOA + βPFOA*AGE)*µX where X is the log transformed values of PFOA for the observations in the age strata, µX is the mean of X, βPFOA, βPFOA*AGE are the coefficient of the term, PFOA and coefficient of the interaction term between PFOA and age respectively. ³ Dashed line represents the mixed effects sensitivity analysis, with a set of observations restricted to those of girls with a observations in the age 12–13 years strata

Figure 3:. Combined effect of PFOA and Age on BMIz in analyses with repeat measurements of serum PFOA, Cincinnati site only

¹The effects of PFOA and other covariates on BMIz were derived from a model which contained race/ethnicity, parent education, PFOA, age at examination, average kilocalories in prior 12 months, hours of moderate or vigorous physical activity in last 12 months ²Age was modeled in four categories (6–7, 8–9, 10–11, 12– 13 and 14 or older) with age 8–9 serving as the comparison group. The combined effect of PFOA and age for each age group was calculated as βPFOA + βPFOA*AGE)*μX where X is the log transformed values of PFOA for the observations in the age strata, μX is the mean of X, βPFOA, βPFOA*AGE are the coefficient of the term, PFOA and coefficient of the interaction term between PFOA and age respectively. ³ The dashed line represents the β effect estimates that resulted from the mixed effects models analysis which include all of the girls and also the repeat measures of serum PFOA. The solid line represents the β effect estimates when only the baseline serum PFOA measurements were included in the model (which is the same as the blue line in Figure 2).