Predictors of orthostatic hypotension in patients attending a transient ischaemic attack clinic: Database study

Figures & data

Figure 1. Distribution of blood pressure (BP) categories for the study subjects. The grey colour shows the proportion of participants in each category with diagnosed hypertension (i.e. T = True, F = False) prior to clinical assessment at the transient ischaemic attack (TIA) clinic.

Table I. Sex specific sample characteristics of the construction sample.

	Males	Females
	n = 1364 (50.6%)	n = 1332 (49.4%)
Age, years	72.0 (64.0–79.0)	75.0 (66.0–81.8)
Supine systolic BP, mmHg	151.0 (132.3–167.0)	154.0 (137.0–171.0)
Median systolic BP standing, mmHg (IQR)	147.0 (131.0–164.0)	149.0 (133.0–167.0)
Median diastolic BP supine, mmHg (IQR)	76.0 (68.0–85.0)	75.0 (66.0–84.0)
Mean diastolic BP standing, mmHg (IQR)	80.0 (72.0–89.0)	79.0 (70.0–88.0)
Mean body mass index, kg/m^2(IQR)	27.0 (24.0–30.0)	26.0 (23.0–30.0)
Waist measurement, cm (IQR)	98.0 (92.0–106.0)	87.0 (79.0–96.0)
Current smoker (%)	196 (14.4)	150 (11.3)
Ex-smoker (%)	789 (57.8)	464 (34.8)
Alcohol (%)	189 (13.9)	80 (6.0)
Previous conditions (%)
Stroke	135 (9.9)	104 (7.8)
TIAs	138 (10.1)	127 (9.5)
Atrial fibrillation	165 (12.1)	138 (10.4)
Hypertension	718 (52.6)	761 (57.1)
Ischaemic heart disease	265 (19.4)	180 (13.5)
Other cardiac disease	59 (4.3)	49 (3.7)
Peripheral vascular disease	58 (4.3)	36 (2.7)
Obesity (≥ 30 kg/m^2)	353 (25.9)	340 (25.5)
Migraine	203(14.9)	361 (27.1)
Diabetes	203(14.9)	148 (11.1)
Known hyperlipidaemia	494 (36.2)	433 (32.5)
Medications (%)
Thiazide	179 (13.1)	295 (22.2)
ACE/ARB	515 (37.8)	428 (32.1)
Beta-blockers	278 (20.4)	255 (19.1)
Diuretic	139 (10.2)	151 (11.3)
Calcium-channel blockers	226 (16.6)	241 (18.1)
Alpha-blocker	81 (5.9)	63 (4.7)
Spirolactone	1 (0.1)	1 (0.1)
Other therapy	61 (4.5)	84 (6.3)
Diagnosis
Non-TIA	558 (40.9)	620 (46.6)
TIA	806 (59.1)	712 (53.4)

Data presented are median (interquartile range, IQR) for continuous variables and number (%) for categorical variables. BP, blood pressure; TIA, transient ischaemic attack; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker.

Table II. Prevalence (number (%)) of patients with postural hypotension in construction and validation datasets.

	Males	Females	All
Construction data
Systolic drop (≥ 20 mmHg)	n = 207 (15.2)	n = 244 (18.3)	n = 451 (16.7)
Diastolic drop (≥ 10 mmHg)	125 (9.2)	133 (10.0)	258 (9.6)
Either (OH)	274 (20.1)	327 (24.6)	601 (22.3)
Both	58 (4.3)	50 (3.8)	108 (4.0)
Validation data
Systolic drop (≥ 20 mmHg)	n = 13 (11.1)	n = 21 (14.7)	n = 34 (13.1)
Diastolic drop (≥ 10 mmHg)	10 (8.6)	8 (5.6)	18 (6.9)
Either (OH)	17 (14.5)	24 (16.8)	41 (15.8)
Both	6 (5.1)	5 (3.5)	11 (4.2)

OH, orthostatic hypotension.

Table III. Summary of univariate analysis.

BP drop	Associated variables	p-values
Systolic (≥ 20 mmHg)	Previous TIA	< 0.005
	Not obese	< 0.005
	Diabetes	< 0.005
	Higher PP	< 0.001
	Higher MAP	< 0.001
	Older age	< 0.001
	Female sex	< 0.05
	Ex-smoker	< 0.05
	Lower BMI	< 0.05
	Smaller waist size	< 0.05
Diastolic (≥ 10 mmHg)	Diabetes	< 0.005
	Previous hypertension	< 0.005
	Higher MAP	< 0.001
	Older age	< 0.001
	Previous TIA	< 0.05
	Previous IHD	< 0.05
	Atrial fibrillation	< 0.05
	ACEi/AIIRBs	< 0.05
	beta-blockers	< 0.05
	Number of medications	< 0.05

BP, blood pressure; TIA, transient ischaemic attack; PP, pulse pressure; MAP, mean arterial pressure; BMI, body mass index; IHD, ischaemic heart disease; ACEi, angiotensin-converting enzyme inhibitor; AIIRB, angiotensin II receptor blocker.

Table IV. Parameter estimates for logistic regression models.

	Odds ratio	95% CI		B	SE	p-value
Systolic drop model
Intercept				− 4.266	0.641	0.000
Age	1.011	0.999	1.023	0.011	0.006	0.063
[TIAs = T]	1.686	1.233	2.306	0.523	0.160	0.001
BMI	0.962	0.938	0.986	−0.039	0.013	0.002
[Diabetes = T]	1.537	1.147	2.060	0.430	0.149	0.004
[Current smoker = T]	1.620	1.148	2.288	0.483	0.176	0.006
[Ex smoker = T]	1.396	1.108	1.759	0.333	0.118	0.005
[Beta-blocker – before = No]	1.288	0.979	1.695	0.253	0.140	0.070
Pulse pressure	1.030	1.024	1.036	0.029	0.003	0.000
Diastolic drop model
Intercept				− 12.148	0.821	0.000
Age	1.052	1.036	1.069	0.051	0.008	0.000
[TIAs = T]	1.480	0.967	2.265	0.392	0.217	0.071
[Diabetes = T]	2.682	1.860	3.869	0.987	0.187	0.000
[Current smoker = T]	1.533	0.976	2.407	0.427	0.230	0.064
[Alcohol = T]	0.567	0.315	1.020	− 0.568	0.300	0.058
[Beta-blocker – before = No]	0.685	0.487	0.964	− 0.379	0.174	0.030
MAP	1.090	1.078	1.102	0.086	0.006	0.000
Pulse pressure	0.961	0.953	0.969	− 0.040	0.004	0.000
OH model
Intercept				− 7.661	0.644	0.000
Age	1.026	1.015	1.036	0.025	0.005	0.000
[Sex – male = T]	0.755	0.615	0.927	− 0.281	0.105	0.007
[Previous stroke = T]	1.317	0.945	1.837	0.276	0.170	0.104
[TIAs = T]	1.580	1.167	2.140	0.458	0.155	0.003
BMI	0.976	0.954	0.998	− 0.024	0.011	0.035
[Diabetes = T]	1.811	1.377	2.382	0.594	0.140	0.000
[Current smoker = T]	1.514	1.091	2.101	0.414	0.167	0.013
[Ex smoker = T]	1.328	1.065	1.656	0.284	0.113	0.012
MAP	1.049	1.042	1.055	0.048	0.003	0.000

The reference category in each model is the FALSE category. CI, confidence interval; TIA, transient ischaemic attack; BMI, body mass index; MAP, mean arterial pressure; OH, orthostatic hypotension.

Table V. C5.0 Rule set containing two rules for orthostatic hypotension (OH) True and two rules for OH False categories.

Rules for OH = True – contains 2 rule(s)

if Peripheral vascular disease = T and MAP > 111 then T

(22 instances; 77.27%)

if MAP > 132.7 then T (116 instances; 67.2%)

Rules for OH = False – contains 2 rule(s)

if MAP < = 111 then F (2021instances; 83.7%)

if Peripheral vascular disease = F and MAP < = 132.7 then F

(2490 instances; 79.9%)

Default: False

The reference category is the false category. The figures in parenthesis show the number of instances that each rule applies to and the confidence of the rule. Rules are not mutually exclusive. This is only one example. MAP, mean arterial pressure.