Table 5.

Reproductive duration and P-wave duration and dispersion by hormone use status in n=31 538* women in the Women’s Health Initiative Clinical Trial

Dependent variable	Hormone use status	Unadjusted effect (95% CI)	Adjusted effect (95% CI)	P values
Due to the fact that there was statistically significant effect modification by HT use on the association between reproductive period and P-wave duration in linear regression models, we present the model estimates by strata of HT use
P-wave duration (ms)	Never user	0.07 (0.03 to 0.11)	0.09 (0.06 to 0.13)	P value for interaction= 0.0009
	Past	−0.04 (−0.08 to 0.005)	0.01 (−0.03 to 0.05)
	Current	−0.03 (−0.06 to 0.004)	0.01 (−0.02 to 0.05)
P-wave dispersion (ms)	Never user	0.002 (−0.04 to 0.05)	0.01 (−0.03 to 0.06)	P value for interaction= 0.65
	Past	−0.03 (−0.09 to 0.02)	−0.01 (−0.06 to 0.05)
	Current	−0.04 (−0.08 to 0.003)	−0.02 (−0.06 to 0.03)

Effect estimates correspond to expected milliseconds increase in PR measure. These models contained an interaction term for reproductive period duration hormone use status. Fully adjusted models were adjusted for number of live births, age, baseline BMI, baseline hypertension status, history of diabetes, income, education, race/ethnicity, region, history of breast feeding, duration of breast feeding, antianxiety medication, antidepressant medication, lipid medication, oophorectomy status, hysterectomy status, hormone use history, heart rate and QRS duration.

*n differs from main analyses due to the exclusion of women with implausible PR-wave measures.

BMI, body mass index; HT, hormone therapy.