A smartwatch sphygmomanometer-based model for predicting short-term new-onset hypertension in individuals with high-normal blood pressure: a cohort study

Figures & data

Figure 1. Flow chart of this study.

Table 1. Baseline characteristics between training and validation dataset.

	Training dataset (n = 3180)				Validation dataset (n = 1000)
Variable	total	HNBP (n = 2487)	New-onset HT (n = 693)	P	total	HNBP (n = 760)	New-onset HT (n = 240)	P	Total P^#
Sex (male, %)	2607(82.0)	2027(81.5)	580(83.7)	0.185	772 (77.2)	584(76.8)	188 (78.3)	0.631	<0.001
Age (IQR, years)	47.0(31.0, 56.0)	37.0(29.0, 53.0)	60.0(45.0, 66.0)	<0.001	47.0(37.0, 57.0)	45.5(35.0, 53.0)	62.0(44.0, 68.0)	<0.001	<0.001
BMI (IQR, kg/m^2)	24.9(21.8, 26.3)	24.4(21.1, 25.6)	28.3(26.0, 31.1)	<0.001	24.7(21.9, 26.3)	24.3(21.3, 25.5)	27.5(25.4, 29.5)	<0.001	0.637
Region (North, %)	1600(50.3)	1138(45.7)	462(66.7)	<0.001	422(42.2)	250(32.9)	172(71.7)	<0.001	<0.001
HR (IQR, beats/min)	68.0(58.0, 75.0)	64.0(55.0, 73.0)	81.0(68.0, 90.0)	<0.001	70.0(59.0, 79.0)	65.0(57.0, 75.0)	89.0(69.0, 101.8)	<0.001	<0.001
Sleep duration (%)*	1341(42.2)	945(38.0)	396(57.1)	<0.001	458(45.8)	298 (39.2)	160(66.7)	<0.001	0.043
Sleep time (%)^&	1908(60.0)	1453(58.4)	455(65.6)	<0.001	530(53.0)	365(48.0)	165(68.8)	<0.001	<0.001

*Percentage of sleep duration less than 420 minutes; ^& Percentage of sleep time after 12 AM; ^# Comparison between training and validation dataset.

Table 2. COX regression analysis of the training dataset*

	Univariate COX regression			Multivariate COX regression
Variable	HR	95% CI	P	HR	95% CI	P
Sex^1	1.139	0.931–1.393	0.207	1.044	0.853–1.278	0.675
Age^2	2.995	2.521–3.559	<0.001	2.929	2.445–3.508	<0.001
BMI^3	2.617	2.177–3.146	<0.001	1.633	1.347–1.980	<0.001
Region^4	2.144	1.831–2.511	<0.001	2.358	2.011–2.765	<0.001
HR^5	2.274	1.943–2.662	<0.001	1.827	1.547–2.157	<0.001
Sleep duration^6	1.987	1.709–2.309	<0.001	1.194	1.011–1.409	0.036
Sleep time^7	1.327	1.135–1.553	<0.001	1.320	1.120–1.555	0.001

*Continuous variable was converted to categorical variable;1: male subjects; 2: Age ≥45 years old; 3: BMI ≥24 kg/m²; 4: lived in North China; 5: HR ≥ 70bmp; 6: Sleep duration <420 mins; 7: fall asleep after 24AM.

Figure 2. Nomogram for new-onset hypertension in HNBP populations. To use the nomogram, locate the individual’s value on each variable axis, and draw a line upward to determine the corresponding points. The sum of these points is located on the total points axis, and a line is drawn downwards to the accumulative axes to determine the likelihood of new-onset hypertension within 6 months.

Figure 3. Restricted cubic spline diagram. (a): RCS of age. (b): RCS of BMI. (c): RCS of HR. (d): RCS of sleep duration.

Figure 4. The calibration curve for predicting new-onset hypertension at 6-month in the training set and validation set. (a): the calibration curve for predicting new-onset hypertension at 6 months in the training set. (b): the calibration curve for predicting new-onset hypertension in 6-month in the validation set.

Figure 5. Kaplan-Meier cumulative curve of the whole dataset. Low-risk group: 0–7.81 scores; middle-risk group: 7.82–9.93 scores; high-risk group: ≥9.94 scores.