Lipid-lowering by postmenopausal hormone therapy (HRT) explains only partly
the assumed coronary risk reduction associated with therapy. To explore ot
her possible mechanisms, we studied associations of HRT use with inflammati
on and hemostasis risk markers in women greater than or equal to 65 years o
f age. Subjects were selected from 3393 participants in the fourth year exa
mination of the Cardiovascular Health Study, an observational study of vasc
ular disease risk factors. After excluding women with vascular disease, we
compared levels of inflammation and hemostasis variables in the 230 women u
sing unopposed estrogen and 60 using estrogen/progestin, with those of 196
nonusers selected as controls. Compared with nonusers, unopposed estrogen u
se was associated with 59% higher mean C-reactive protein (P<0.001), but wi
th modestly lower levels of other inflammation indicators, fibrinogen, and
alpha-1 acid glycoprotein (P<0.001). Factor VIIc was 16% higher among estro
gen users (P<0.001), but this was not associated with higher thrombin produ
ction (prothrombin fragment 1-2), or increased fibrin breakdown (D-dimer).
Concentration of plasminogen activator inhibitor-1 was 50% lower in both us
ing groups (P<0.001) compared with nonusers, and this was associated with h
igher plasmin-antiplasmin complex: 8% higher in estrogen and 18% higher in
estrogen/progestin users (P<0.05). Relationships between the markers and ho
rmone use were less pronounced in estrogen/progestin users, with no associa
tion for C-reactive protein except in women in upper 2 tertiles of body mas
s index (P for interaction, 0.02). The direction and strength of the associ
ations of HRT use with inflammation markers differed depending on the prote
in, so it is not clear whether HRT confers coronary risk reduction through
an inflammation-sensitive mechanism. Associations with hemostasis markers i
ndicated no association with evidence of procoagulation and a possible asso
ciation with increased fibrinolytic activity.