No abstract

Our findings indicate a modest positive association of PFOA on some lipid parameters and a need for follow-up studies.

Introduction JThis report summarizes the statistical analysis of serum chemistry, hematology, and thiodiglycol data collected in Phase lB of M1REF Task 94-33. This report should be considered a supplement to the Phase I B statistical report dated February 5, 1998. The conclusions drawn in the previous report remain valid. Statistical Methods Serum chemistry and hematology data for each animal on study days 0, 1, 3, and 7 were included I in the statistical analysis. Twice daily thiodiglycol readings from study days 0-7 were included in the analysis. For each parameter, the following analysis of variance (ANOVA) model was fitted to the data: yjI`+Ot+Tj+Ej where yij is the reading for animal j on study day i, g. is the overall mean, ai is a fixed effect for study day i, rj is a random effect for animal j, and Eij is a random error term. Appropriate contrasts were used to estimate the difference in means between study days. For the serum Schemistry and hematology parameters, all pairwise comparisons between study day means were estimated. For the thiodiglycol parameters, AM and PM means on each study day were compared to the pretreatment mean (study day 0, AM reading). For each parameter, a Bonferroni adjustment for multiple comparisons was applied to ensure that the probability of making at least one incorrect conclusion of significance is no higher than 0.05. The SAS (V6.12) MIXED procedure was used to fit the statistical models. Results Descriptive statistics and statistical comparisons are presented in Tables 1, 2, and 3, for the serum chemistry, hematology, and thiodiglycol parameters, respectively. Figures A-I through A-3 8 present the means and 95 percent confidence intervals for each parameter, in the order they are presented in Tables 1 to 3. Similarly, Figures B-1 through B-38 present the data for each animal. Statistical comparisons of serum chemistry and hematology data are summarized by the letters entered in the last row for each parameter in Tables 1 and 2. For these comparisons, study day means that were not significantly different (at an over all 0.05 level) share at least one letter and those that were significantly different do not have a common letter. When no significant differences were noted among the study days, the single letter A appears in each cell (see Aspartate Transaminase). For parameters where significant differences were noted, the results are summarized below, in addition to the table. D •. D-2 Serum Chemistry Parameters (Table 1) Alanine Transaminase: Means on days 1 and 3 were greater than that on day 7. No differences were noted in comparisons to the mean on day 0. Albumin: The mean on day 3 was greater than that on day 7. Alkaline Phosphatase: The mean on day 0 was greater than on days 1, 3, and 7. The mean on day 1 was greater than on days 3 and 7. The means on days 3 and 7 were not significantly different. Amylase: The mean on day 0 was greater than on days 1, 3, and 7. Blood Urea Nitrogen: The mean on day 0 was less than on days 1, 3, and 7. Calcium : The mean on day I was s...

Challenges to the use of the no observed effect concentration (NOEC) in ecotoxicology have appeared over the years, with a recent call for banning its use in favor of the x% effects concentration (ECx). This article presents an opposing view, providing reasons for the continued use of the NOEC, and for hypothesis testing in general. Although the use of ECx values is appropriate in many situations, there are numerous real-world examples where it is not suitable and offers no advantage over the use of hypothesis testing. These examples are presented with recommended data analysis techniques, illustrating the variety of statistical approaches that are meaningful in analyzing ecotoxicity data. Thoughtful consideration of study design and proper analysis and interpretation of the results will go further to advance the science of ecotoxicology than attempting to implement a blanket prohibition or endorsement of any single statistical approach. Integr Environ Assess Manag 2013;9:12-16. ß 2012 SETAC