Effect of weight loss, with or without exercise, on body composition and sex hormones in postmenopausal women: the SHAPE-2 trial

Table 4 Baseline and 16-week differences in serum sex hormones and treatment effects between study groups

		Baseline geometric mean	Geometric mean at 16 wk	Change at 16 wk, %	TER^a (95 % CI), intervention vs control	P value*	TER^a (95 % CI), mainly exercise vs diet	P value**
Oestradiol (pg/ml)
	Control	3.89	4.01	3.11
	Diet	4.20	3.62	−13.8	0.86 (0.75–0.98)	0.025
	Mainly exercise	3.69	3.22	−12.7	0.83 (0.73–0.95)	0.007	0.97 (0.87–1.08)	0.562
Oestrone (pg/ml)
	Control	20.1	20.4	1.51
	Diet	20.4	20.1	−1.26	0.98 (0.88–1.08)	0.650
	Mainly exercise	19.9	18.5	−6.67	0.92 (0.82–1.02)	0.109	0.94 (0.86–1.02)	0.154
Free oestradiol (pg/ml)
	Control	0.09	0.10	3.23
	Diet	0.10	0.08	−17.7	0.80 (0.70–0.92)	0.002
	Mainly exercise	0.09	0.07	−19.1	0.77 (0.67–0.88)	<0.001	0.96 (0.85–1.07)	0.425
Testosterone (pg/ml)
	Control	194	186	−4.07
	Diet	197	189	−3.76	1.01 (0.92–1.10)	0.886
	Mainly exercise	186	172	−7.63	0.96 (0.87–1.05)	0.332	0.95 (0.88–1.02)	0.166
Androstenedione (pg/ml)
	Control	575	560	−2.60
	Diet	562	537	−4.50	0.97 (0.85–1.12)	0.684
	Mainly exercise	573	488	−14.7	0.87 (0.76–1.00)	0.059	0.90 (0.80–1.01)	0.064
Free testosterone (pg/ml)
	Control	2.71	2.61	−3.90
	Diet	2.53	2.25	−11.2	0.91 (0.83–1.01)	0.069
	Mainly exercise	2.44	2.01	−17.7	0.84 (0.76–0.93)	0.001	0.92 (0.85–0.99)	0.043
SHBG (nmol/L)
	Control	44.2	44.0	−0.30
	Diet	50.7	57.1	12.6	1.14 (1.07–1.23)	<0.001
	Mainly exercise	49.3	58.6	19.0	1.21 (1.12–1.30)	<0.001	1.05 (1.00–1.12)	0.070

Baseline and follow-up measurements of complete cases (i.e., women with both baseline and follow-up measurements) are presented. Complete case data of oestradiol were available for 223 women; oestrone for 221 women; free oestradiol for 222 women; testosterone and androstenedione for 229 women; free testosterone for 228 women; and SHBG for 230 women
Abbreviations: CI confidence interval, SHBG sex hormone-binding hormone, TER treatment effect ratio
*P < 0.025 was considered significant for the comparison of both intervention groups vs control
**P < 0.05 was considered significant for the comparison mainly exercise vs diet
^aTER represents the overall intervention effect on hormone change (adjusted for baseline), estimated by linear regression analysis. Because the linear regression models were based on log-transformed hormone data, the presented treatment effect is the antilogarithm of the original estimate. Therefore, the TER is a ratio that indicates how many times the level in one group is higher (TER >1) or lower (TER <1) than a reference group. For example, TER intervention vs control of 0.9 indicates that the hormone level in the intervention group is, on average, 10 % lower than in the control group

ISSN: 1465-542X