Table 3.

Risk of Bias ratings of each unit of evidence for all the questions involved in the risk-of-bias analysis for studies with two or more chemical doses.

Risk of Bias questions	++ Definitely low risk of bias	+ Probably low risk of bias	~ Probably high risk of bias	~~ Definitely high risk of bias
					KEY ELEMENTS
1. Were reliable and sensitive methods used for the T4 quantification?	1	27	31	0
2. Were reliable and sensitive methods used for the TSH quantification?	15	16	28	0
3. Was exposure sufficiently characterised?	33	26	0	0
4. Was the number of animals per dose group sufficient?	39	12	8	0
5. Was a positive control included?	1	56	2	0
6. Were measurements collected at a suitable timepoint when we compare control and treated groups?	7	50	2	0
7. Have the authors evaluated the hormonal changes in relation to contemporaneous or historical controls?	59	–	–	0
8. Was the diet free of phytoestrogens or/and goitrogens?	10	31	–	18
OTHER ELEMENTS
Were animals randomly allocated to dose groups?	29	30	0	0
Was allocation to dose groups adequately concealed?	0	59	0	0
Were all experimental animals of similar age, strain, health status and source?	55	4	0	0
Were experimental conditions identical across study groups?	50	9	0	0
Were research personnel blinded to study groups?	0	59	0	0
Were outcome data complete without attrition or exclusion?	14	45	0	0
Were the factors that might influence the variability of TH measurements considered?	8	51	0	0
Have all study outcomes been reported?	23	35	0	1
Have funding sources and conflicts of interest been reported?	34	–	–	25
Were statistical methods reported in the study, appropriate and consistent?	14	44	1	0

Shadings are related to number of items, the higher the number, the darker the shading.