Table 3.

Multivariable log-binomial regression for the association between body mass index and new-onset of heart failure among Thai patients with hypertension, stratified by blood pressure control for the latest two consecutive visits

Variables	Control BP		Uncontrol BP		p for interaction
	aRR (95% CI)*	p-value	aRR (95% CI)*	p-value
Body mass index (BMI)
Continuous (linear)	1.10 (1.03–1.17)	0.007	1.09 (1.04–1.14)	< 0.001	0.477
Continuous (quadratic)	1.002 (1.001–1.002)	0.010	1.001 (1.001–1.002)	< 0.001	0.763
BMI category, kg/m2					0.602
< 18.50	0.55 (0.12–2.57)	0.447	1.62 (0.41–6.32)	0.490
18.50–22.99	Ref.		Ref.
23.00–24.99	1.22 (0.41–3.68)	0.720	1.04 (0.33–3.29)	0.948
25.00–29.99	1.61 (0.62–4.17)	0.327	1.53 (0.59–3.99)	0.382
≥ 30.00	4.85 (1.77–13.28)	0.002	3.21 (1.17–8.80)	0.024
BMI ≥ 25 kg/m2					0.642
No	Ref.		Ref.
Yes	2.33 (1.12–4.83)	0.023	1.83 (0.93–3.58)	0.080
BMI ≥ 30 kg/m2					0.926
No	Ref.		Ref.
Yes	4.05 (1.73–9.51)	0.001	2.52 (1.25–5.06)	0.010

BMI body mass index, BP blood pressure, aRR adjusted Risk Ratio, 95% CI 95% confidence interval

^*Multivariable analysis: adjusting for age, sex, health schemes, geographic regions, location of outpatient clinic, type 2 diabetes, dyslipidemia, renal insufficiency, history of ischemic heart disease, smoking status, alcohol use, control blood pressure consecutively two latest visits, and ACEI/ARB use