We observed no mortality in the experiments in this study. The two higher mixture dosages (i.e., 76.2 and 152.4 mg/kg) evoked few clinical signs of pyrethroid toxicity. Mild whole-body tremors were present in most animals at the highest mixture dose (152.4 mg/kg) from 1 to 4 hr postdosing. Signs of high-dose pyrethroid toxicity such as excessive salivation, whole-body tremors, and choreoathetotic movements (Aldridge 1990
; McDaniel and Moser 1993
; Soderlund et al. 2002
) were not observed in any animals, with one exception: One rat in the time-course experiment exhibited clinical signs of excessive pyrethroid exposure and raspy breath sounds, possibly due to partial aspiration of the gavage solution into the lungs. Data from this animal were not used.
The time-course study revealed a rapid decline in motor activity with a peak decrease at 1–2 hr postdosing (). Activity recovered to control levels at 24–48 hr postdosing. The time course of effects was similar for both mixture-dose groups. These conclusions were supported by a significant mixture-dose × time interaction [F(10,147) = 5.18, p < 0.0001] and significant main effects of mixture dose [F(2,147) = 62.48, p < 0.0001] and testing time [F(4,147) = 15.5, p < 0.0001]. The activity of all mixture-dose groups was significantly decreased compared with controls at 1, 2, 4, and 8 hr (p< 0.05).
Time course of cumulative effects of 11 pyrethroids on figure-eight maze activity (mean ± SE). The arrow indicates the time of peak effects for the tested mixture.
Parameter estimates and corresponding p-values from the additivity model in Equation 2a, from the simultaneously fit single-chemical data and mixture data, are provided in . The slope parameters associated with each of the 11 single chemicals (β values) and for the fixed-ratio mixtures (θ values) were negative and significant, indicating that as the dose of each individual chemical or the total dose of the mixture increases, the mean motor activity decreases.
Estimated model parameters from the threshold additivity model (Equation 2a) and the mixture model (Equation 3a) fit simultaneously.
shows the plots of individual data points for each chemical and the fit dose response using the additivity model. These plots illustrate the wide potency range of the individual chemicals, from 10 to 900 mg/kg, and the dose-related decrease in activity for all 11 chemicals. These plots also illustrate the estimated thresholds for each of the dose–response functions. lists the estimated model parameters from the threshold additivity model and the mixture model fit simultaneously. The simultaneous fit of the additivity model and the mixture model accommodates a common maximum effect parameter (α), which was estimated to be 26.5 (i.e., 26.5% of control; ). The γ parameter was constrained to be γ = 100 – α, so that the mean response for the control groups is 100%. and list the model estimates for the threshold dose and ED30
dose. Note that the individual chemical estimates in – are in some cases marginally different than previously published estimates (Wolansky et al. 2006
) because of the inclusion of the mixtures data in the SCR model in the present study: The empirical mixtures data were not available for inclusion at the time the previous estimates were published.
Observed data (individual data points) and the model-predicted dose–response curve from the additivity threshold model given in Equation 2a for each of the 11 pyrethroids.
Estimated dose thresholds for each of the 11 chemicals with 95% large sample CIs.
Estimated ED30 values for each of the 11 chemicals with 95% large-sample CIs.
The mixture dose–response studies demonstrated dose-related decreases in activity regardless of the dosing protocol (). For the SQT group, the pyrethroid mixture decreased activity by approximately 58% in the two highest mixture-dose groups, with significant decreases in all mixture doses ≥ 30.4 mg/kg (p < 0.05). There was no interaction between testing block and mixture dose [F(14,92) = 0.66, p < 0.8030], but there were significant main effects of dose [F(7,92) = 12.84, p < 0.0001] and block [F(2,92) = 6.15, p < 0.0035]. The main effect of block was due to slight differences in the overall baseline for activity counts over the different test days (data not shown). For the SLT protocol, the pyrethroid mixture decreased activity by 60% in the highest mixture-dose group, with significant decreases in all mixture doses ≥ 50.2 mg/kg (p < 0.05). There was no interaction between testing block and mixture dose [F(14,94) = 1.24, p < 0.2643], but there were significant main effects of dose [F(7,94) = 20.45, p < 0.0001] and block [F(2,94) = 2.85, p < 0.0414]. Because of the lack of interaction between replicate testing blocks and mixture dose, and the significant main effect of replicate block, we conducted all additional analyses on the motor activity counts expressed as percentage of block control values.
Figure 3 Dose–response relationships for the cumulative effects of 11 pyrethroids on figure-eight maze activity (mean ± SD). (A) SLT group. (B) SQT group. The departure of the experimental data from the predictive curve modeled assuming dose addition (more ...)
Results of the SCR method demonstrated no significant difference between the predicted response and the empirical data for both the SLT and SQT exposure protocols. lists the slope estimates for individual chemicals and the dosing protocols, as well as the threshold estimates for both protocols. The empirical fit for the mixture administered using the SLT protocol was not different from that predicted assuming additivity, and the null hypothesis was not rejected [F(2,1037) = 0.015, p = 0.985; ]. The small shift to the left in the dose–response relationship between the empirical and predicted curves for the SQT protocol () was not significant, and the null hypothesis was not rejected [F(2,1037) = 2.65, p = 0.071]. The two thresholds were not statistically different; however, the SQT protocol threshold value (5.38 mg/kg) was 3.7-fold lower than that using the SLT protocol (19.82 mg/kg). This difference was 1.5-fold when we compared mixture ED30 values (SQT, 29.27 mg/kg, vs. SLT, 49.81 mg/kg). The CIs were wide and included zero, and although the threshold for the mixture administered using the SQT dosing protocol was numerically lower than that for the mixture where the chemicals were administered together at once (see ), a test of coincidence in the two mixture curves was not rejected [F(2,1037)= 0.90, p= 0.407; ].