Cardiac vagal control and theoretical models of co-occurring depression and anxiety: A cross-sectional psychophysiological study of community elderly

Table 3 Multiple regression analysis on the association of cardiac vagal control and the Hospital Anxiety and Depression Scale ^†

	High frequency component of heart rate variability [ln(ms²)]
Factor models	β (SE)	P-value
Model I. Unitary factor model of Razavi (1999)
Total score of full scale	−0.01 (0.01)	0.46
Model II . Original two-factor model of Zigmond & Snaith (1983)
Anhedonic depression^‡	−0.21 (0.09)	0.02
Anxiety subscale	0.02 (0.04)	0.54
Model III. Flat tripartite model of Dunbar et al. (2000)
Anhedonic depression	−0.22 (0.09)	0.01
Negative affectivity	−0.09 (0.05)	0.12
Autonomic anxiety	0.24 (0.09)	0.01
Model IV. Hierarchical tripartite model of Dunbar et al. (2000)
Direct effect
Anhedonic depression	−0.22 (0.09)	0.01
Negative affectivity	−0.09 (0.05)	0.12
Autonomic anxiety	0.24 (0.09)	0.01
Indirect effect	β (SE)	BCa 95% CI^§
Negative affectivity → Anhedonic depression	−0.05 (0.03)	(−0.11, -0.01)
Negative affectivity → Autonomic anxiety	0.08 (0.03)	(0.02, 0.15)

†Variables forced to enter into the regression models include: Age (years), Education (≧13, 10–12, 7–9, <7 years), Diabetes mellitus (yes/no), Hypertension (yes/no), Cardiovascular disease (yes/no), Frequency of exercise (≧ 1 vs <1/week), Current smoker (yes/no) , Site of data collection (Hospital vs home), Body mass index (≧ 25 , vs <25 kg/m²), and various combination of factor items.
‡Total score of depression subscale was rescaled by quartiles as <5 (lowest 2 quartiles), 5–6 and ≧7.
§BCa: bias corrected and accelerated confidence interval derived from 5000 bootstrap samples.

ISSN: 1471-244X