Table 4.

Predictors of FT3 in multivariate linear regression analysis with FT3 as dependent variable and sex, T1DM duration, HbA_1c, WHR as independent variables. Model performance: R² = 0.15, p < 0.0001

Predictors	Multivariate analysis
	β	p-value
Male sex	0.16	0.02
T1DM duration	−0.15	0.01
HbA1c	−0.25	<0.0001
WHR	0.08	0.22

Following predictors were taken into account: T1DM duration, TSH, FT4, BMI, WHR, HbA_1c and sex. Predictors were chosen based primarily on statistically significant correlations (Table 3). Sex was inserted into regression based on widely known dependence of hypothyroidism on sex. BMI was not inserted into the analysis because of significant correlation with WHR and WHR was chosen because this predictor was better correlated with FT3. FT4 was inserted into second regression model (Table 5) to assess predictors from this model with full model. TSH was not inserted into analysis because significant correlation with FT3 did not occur

Bold p-values are statistically significant (a p-value less than 0.05)

Abbreviations: see Table 2