Depressive symptoms and antidepressant use in relation to white blood cell count among postmenopausal women from the Women's Health Initiative.

Inflammation can play a role in the pathophysiology of depression, and specific types of antidepressants may have inflammatory or anti-inflammatory properties. Furthermore, depression and antidepressant use has been linked to white blood cell (WBC) count, a routinely measured inflammatory marker. We examined the cross-sectional and longitudinal relationships of depressive symptoms and/or antidepressant use with WBC count among postmenopausal women. Analyses of cross-sectional data at enrollment were performed on 125,307 participants, 50-79 years of age, from the Women's Health Initiative Clinical Trials and Observational Studies who met eligibility criteria, and a subset of those with 3-year follow-up data were examined for longitudinal relationships. Depressive symptoms were defined using the Burnam Algorithm whereas antidepressant use was defined using therapeutic class codes. WBC count (Kcell/ml) was obtained through laboratory evaluations of fasting blood samples. Multivariable regression modeling was performed taking sociodemographic, lifestyle and health characteristics into consideration. At enrollment, nearly 85% were non-users of antidepressants with no depressive symptoms, 5% were antidepressant users with no depressive symptoms, 9% were non-users of antidepressants with depressive symptoms, and 2% were users of antidepressants with depressive symptoms. In fully-adjusted models, cross-sectional relationships were observed whereby women in the 2 (OR = 1.06, 95% CI: 1.01, 1.13), 3 (OR = 1.06, 95% CI: 1.00, 1.12) or 4 (OR = 1.10, 95% CI: 1.05, 1.17) quartiles of WBC count were more likely to exhibit depressive symptoms, and women in the 4 quartile were more likely to be users of antidepressants (OR = 1.07, 95% CI: 1.00, 1.15), compared to women in the 1 quartile. Compared to women who exhibited no depressive symptoms at either visit, those with consistent depressive symptoms at enrollment and at 3-year follow-up had faster decline in WBC count (β = -0.73, 95% CI: -1.33, -0.14) over time. No significant bidirectional relationships were observed between changes in depressive symptoms score and WBC count over time. In conclusion, depressive symptoms and/or antidepressant use were cross-sectionally related to higher WBC counts among postmenopausal women. Further evaluation of observed relationships is needed in the context of prospective cohort studies involving older adult men and women, with repeated measures of depression, antidepressant use, and WBC count.