The Anglo-Scandinavian Cardiac Outcomes Trial –

Blood Pressure Lowering Arm (ASCOT-BPLA)

Blood Pressure Variability and Cardiovascular Outcomes

PS Sever, PM Rothwell, SC Howard, JE Dobson, B Dahlöf,

H Wedel,, NR Poulter,, for the ASCOT Investigators
g
International Centre for Circulatory Health, Imperial College London
and
Stroke Prevention Research Unit, University of Oxford
Treatment algorithm to BP targets < 140/90 mmHg
or < 130/80 mmHg in patients with diabetes

amlodipine 5-10 mg atenolol 50-100 mg

add add

bendroflumethiazide-K
perindopril
p p 4-8 mg
g
1 25 2 5 mg
1.25-2.5

add

doxazosin GITS 4-8 mg

add
additional drugs, eg, moxonidine/spironolactone

Median follow up was for 5.5 years

ASCOT-Blood pressure variability: methods
(
(based on over 1 million BP readings))
• Of 19
19,257
257 patients
patients, 18
18,530
530 had ≥ 2 follow
follow-up
up visits
(median = 10) from 6 months onwards until the end of
tthe
e ttrial
a
• 3 blood pressure measurements were recorded at
each visit, usingg standardised techniques,
q at 6
monthly intervals for a median follow up of 5.5 years
• From 6 months onwards there were 350 strokes and
704 coronary events (non-fatal MI, fatal CHD, new
onset angina, non-fatal and fatal heart failure) in the
atenolol based group and 279 and 611 respectively in
atenolol-based
the amlodipine-based group
 Stroke and coronary risk expressed by decile of measure of visit-to-visit SBP variability
Stroke Risk Coronary Risk

Standard deviation of SBP Amlodipine

Atenolol

Coefficient of variation of SBP

Variation independent of mean

SBP

Decile of measure Decile of measure

 Hazard ratios (95% CI) for the effect of treatment
(amlodipine versus atenolol) on risk of stroke
Parameters calculated using all BP measurements from 6 months onwards. Mean, SD, CV,
and VIM are entered into the model as deciles

Stroke Systolic blood pressure

Variables in model HR (95% CI) p value

Treatment (Rx) 0.78 (0.67–0.90) 0.001

Usual BP
Rx + mean 0.84 (0.72–0.98)
(0.72 0.98) 0.025
Visit-to-visit BP variability
Rx + mean + SD 0.96 (0.82–1.12) 0.59
Rx + mean + CV 0 95 (0
0.95 (0.82
82–1
1.11)
11) 0 55
0.55
Rx + mean + VIM 0.96 (0.82–1.12) 0.58
Within-visit and visit-to-visit BP variability
Rx + within
within-visit
visit SD 0 84 (0
0.84 (0.72
72–0
0.98)
98) 0 024
0.024
Rx + mean + VIM + WVSD 0.99 (0.85–1.16) 0.89
SD, standard deviation; CV, coefficient of variation; VIM, variability independent of mean;
WVSD, within-visit standard deviation
 Hazard ratios (95% CI) for the effect of treatment
(amlodipine versus atenolol) on risk of coronary events
Parameters calculated using all BP measurements from 6 months onwards. Mean, SD, CV,
and VIM are entered into the model as deciles

Coronary Events Systolic blood pressure

Variables in model HR (95% CI) p value

Treatment (Rx) 0.85 (0.77–0.94) 0.002

Usual BP
Rx + mean 0.88 (0.80–0.98)
(0.80 0.98) 0.019
Visit-to-visit BP variability
Rx + mean + SD 1.00 (0.90–1.11) 0.98
Rx + mean + CV 1 00 (0
1.00 (0.90
90–1
1.11)
11) 0 99
0.99
Rx + mean + VIM 1.00 (0.90–1.10) 0.99
Within-visit and visit-to-visit BP variability
Rx + within
within-visit
visit SD 0 88 (0
0.88 (0.79
79–0
0.97)
97) 0 013
0.013
Rx + mean + VIM + WVSD 1.01 (0.91–1.12) 0.88
SD, standard deviation; CV, coefficient of variation; VIM, variability independent of mean;
WVSD, within-visit standard deviation
Ambulatory blood pressure monitoring

• 1905 patients had an average of 3.25 recordings from 6 months

onwards
• Daytime SBP slightly higher but night-time SBP slightly lower
on amlodipine-based
amlodipine based treatment
• Morning surge similar on both treatments and only weakly
correlated
co e ated with
t BP visit-to-visit
s t to s t variability
a ab ty
• Intra ABPM coefficient of variation of SBP correlated with
visit-to-visit variability in clinic SBP
– atenolol group, r = 0.38
– amlodipine
p g
group,
p r = 0.29, p < 0.0001 for both g
groups
p
• Intra ABPM variability in daytime SBP predicted both stroke and
coronary events (but less so than visit-to-visit variability)
 Summary
• Mean BP in trial has minimal effect on stroke outcome and no
effect on CHD outcome
• Various measures of visit-to-visit BP variability (SD, coefficient
of variation and variation independent of mean BP) are powerful
predictors of both stroke and CHD outcomes
• Other measures of variability (within-visit variability and
variability
i bilit assessed
dbby ABPM) also l predict
di t cardiovascular
di l
outcomes but less than visit-to-visit variability
• Amlodipine reduces variability compared with atenolol
• Variability increased with age, diabetes, smoking, and in those
with established vascular disease
• Adjusting for BP variability completely explains differences in
stroke and CHD outcomes between amlodipine-based and
atenolol-based
t l lb d ttreatment
t t in
i ASCOT