Transcript Effect of first TIA on cardiovascular prevention: cohort

Update on hypertension - diagnosis,
monitoring and guideline treatment targets
Prof. Richard McManus, Birmingham, United
Kingdom
Overview

Background
 Routine measurement of blood pressure
(is rubbish)
 Diagnosis of hypertension - ABPM?
 Management of hypertension – Home ?
 Treatment targets – any changes?
 Conclusions
Stroke Risk increases with age & usual BP
High
60-69
Similarly
40-49 for Heart
Disease
Risk
Low
Low
High
Bottom line BP vs Risk
10 mmHg
38% stroke risk
18% CHD risk
If you live long enough you get hypertension
The population is ageing
In mid-2008 the median age of the population was 39 years, up from 37 in 1998.
Routine measurement is often flawed
Dotplot of Last_practice_systolic
Dotplot of systolic
90
100
110
120
130 140 150 160
Last_practice_systolic
170
180
190
200
Each symbol represents up to 4 observations.
Same population with
routine and research
measurement
80
90
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240
systolic
Each symbol represents up to 12 observations.
Blood Pressure varies through the day and
between seasons
Hypertension. 2006;47:155-161
Multiple measurements better estimate
mean blood pressure
This variability means that measurement
error can drown out the truth
Even on a single occasion BP drops
Interval Plot of systolic vs occasion
95% CI for the Mean



144
142
140
systolic

Approx 1500
patients
24 practices
6 readings at 1min
intervals
12 mmHg systolic
drop
Stable after 5th
reading
138
136
134
132
130
1
2
3
4
5
6
occasion
Interval Plot of diastolic vs occasions
95% CI for the Mean
83
82
81
diastolic

146
80
79
78
77
1
2
3
4
occasions
5
6
Family Practice 1997; 14:130-135
BP takes some
time to settle with
repeated
measurement
over
weeks/months
Many factors affect BP measurement
BMJ 2001;322;908-911
Diagnosing hypertension

Traditionally based on clinic measurement
 Most outcome trials use clinic measures
 But
– Flawed measure (one off from continuum)
– Takes weeks / months to make diagnosis
What about ABPM?
 Half
hourly measurements during the day
 Better
 Hourly
 Main
measure usual BP
at night
outcome is mean day time ABPM
 Other
info available (dipping etc)
International Thresholds for hypertension
diagnosis (clinic and ABPM)
Mean
daytime BP
UK (ABPM) = 135/85 mmHg
What’s normal for ABPM (and home)?
Based on Head et al BMJ 2010

adjust by 5/5 mmHg at lower threshold
(stage 1 hypertension, 140/90 mmHg clinic)
– ie < 135/85 mm Hg

10/5 mmHg at higher threshold
(stage 2 hypertension, 160/100 mmHg clinic)
– Ie < 150/95 mmHg
How do clinic and ABPM compare?

Reviewed literature: 2914 studies of which 20
were relevant

7 compared ABPM with clinic monitoring for
diagnosis
 Full details:
BMJ 2011;342:d3621 doi: 10.1136/bmj.d3621
Many people currently potentially
misdiagnosed...
Worse if only studies around
diagnostic threshold used:
sensitivity of 86% and
specificity of 46%
What about Home Monitoring?
Relative
sensitivity and
specificity of
clinic and home
measurement vs
ABPM
Better correlation with end organ
damage and outcome (ABPM)
1963 patients
Mean FU 5 yrs
Baseline ABPM
CVD events
Better correlation with end organ
damage and outcome (Home)
1700 patients, 10 years FU, 150 CVAs
Screening = 2 clinic measurements one occasion
Home = 25 measurements over 4 weeks
Journal of Hypertension 2004, 22:1099–1104
Detection of white coat and masked HT
But what about costs?

Treatment – ↓drug costs

Follow up – ↓clinician costs

But do additional costs of ABPM out weigh
these?
Is ABPM cost effective?

Modelling to evaluate the most cost-effective
method of confirming a diagnosis of
hypertension in a population suspected of
having hypertension

ABPM vs Home vs clinic

Further details Lovibond et al, Lancet 2011
Markov Model

Health service perspective
 Lifetime horizon
 Assume all have raised clinic screening
 People aged 40 and over
Markov Model

Costs from published sources and NHS
 Test performance from systematic review
 Risk calculated using Framingham equation
Results

ABPM most cost effective for every age group
Results

ABPM most cost effective for every age group

Robust to wide range of sensitivity analyses
Results – sensitivity analysis
Results

ABPM most cost effective for every age group

Robust to wide range of sensitivity analyses

Sensitive to
– Assumption of equal test performance
– Assumption of no effect of Rx below
140/90 mmHg
ABPM

Don’t forget ABPM need to be validated and
have yearly calibration (bhsoc.org.uk website)

Lack of night time dipping is additional risk
(hence rationale for night readings)

Currently limited in PC as most practices
either need to refer or only have one ABPM
machine

Commissioners need to consider whole
health economy
Self Monitoring reduces BP

Small reductions in
blood pressure from
self-monitoring:
– SBP by 3.8 mmHg
– DBP by 1.5 mmHg
Bray et al. Annals of Medicine 2010
Self monitoring costs equivalent to usual care
BMJ 2005;331;493
How many measurements?
Conclusion = at least 4 days monitoring and discard 1st
European (& UK) Guideline is 1 week, 2 readings bd,
discard day 1, take mean (limited rationale)
What is the place of home
monitoring?

Management after diagnosis, especially if
proven significant white coat effect
 More outcome and test performance data
needed for diagnosis
 Adjunct to other co-interventions and self
management...
Co-interventions enhance self
monitoring effect
Weighted Mean diff.
(95% CI)
Study
% Weight
Mehos (2000)
-10.10 (-20.61,0.41)
5.4
Green b (2008)
-9.30 (-11.80,-6.80)
15.3
Zillich (2005)
-4.40 (-10.52,1.72)
9.9
Broege(2001)
-2.00 (-16.33,12.33)
3.4
Artinian (2001)
-25.60 (-41.78,-9.42)
2.8
Rudd (2004)
-8.50 (-14.16,-2.84)
10.5
Green a (2008)
-3.40 (-5.91,-0.89)
15.2
Parati (2009)
-0.20 (-3.84,3.44)
13.6
Baque (2005)
-0.14 (-2.05,1.77)
15.3
10.8
Bailey (1999)
5.00 (-6.07,16.07)
2.5
12.9
Verberk (2007)
0.50 (-3.65,4.65)
9.8
Binstock (1988)
-18.00 (-27.13,-8.87)
McManus (2005)
-2.30 (-5.47,0.87)
12.1
Marquez-Contreras (2006)
-4.60 (-9.01,-0.19)
9.2
Midanik (1991)
-2.60 (-7.26,2.06)
8.7
Soghikan (1992)
-3.30 (-6.77,0.17)
11.3
Mulhauser (1993)
-5.00 (-10.45,0.45)
Freidman (1996)
-0.90 (-4.98,3.18)
Overall (95% CI)
-5.29 (-8.26,-2.32)
-30
-15
0
15
30
Weighted Mean diff.
5.3 mmHg
2.5 mmHg
Weighted Mean diff.
(95% CI)
Study
% Weight
3.5
Vetter (2000)
-0.50 (-3.07,2.07)
13.6
Halme (2005)
-3.10 (-7.93,1.73)
8.4
Carnaham (1975)
-7.50 (-14.28,-0.72)
5.5
Overall (95% CI)
-2.52 (-4.43,-0.61)
-30
-15
0
15
30
Weighted Mean diff.
What’s a co-intervention?

Nurses

Telemonitoring

Patient Education

Self Management
Theoretical basis for self management
Patients


Increased patient involvement in management
decisions will result in:
Cues to action
Adherence
Increased self efficacy
Behaviour change
Better use of medication likely to have most effect
Professionals


Systematic titration of medication effective
Evidence of clinical inertia
TASMINH2 Research Questions
Does self management with telemonitoring and
titration of antihypertensive medication by
people with poorly controlled treated
hypertension result in:
1. Better control of blood pressure?
2. Changes in reported adverse events or
health behaviours or costs?
3. Is it achievable in routine practice and is it
acceptable to patients?
The Trial

Eligibility
–
–
–
–

Age 35-85
Treated hypertension (no more than 2 BP meds)
Baseline BP >140/90 mmHg
Willing to self monitor and self titrate medication
Patients individually randomised to
self-management vs usual care
stratified by practice and minimised on sex,
baseline SBP, DM status,
 Practice GPs determine management
Intervention

Self Monitoring – 1st week of every month
Intervention

Blood Pressure Targets:
– NICE (140/90 or 140/80
mmHg)
– minus 10/5 mmHg
i.e. 130/85 mmHg or
130/75 mmHg
 Patients agreed titration
schedule with their GP after
randomisation
 Traffic Light system to adjust
medication
Outcomes

Follow up at 6 & 12 months
 Main outcome Systolic Blood Pressure
 Secondary outcomes: Diastolic BP / costs /
anxiety / health behaviours/ patient
preferences / systems impact
 Recruitment target 480 patients (240 x 2)
 Sufficient to detect 5mmHg difference
between groups
Invited (n = 7637)
Declined Invitation
(n = 5987)
Results
Assessed for eligibility (n = 1650)
Excluded (n = 1123)
Not Eligible (n = 1044)
Declined to participate (n=79)
Randomised (n = 527)
Control (n = 264)
Received usual care
(n = 264)
Did not attend follow up
(n=14)*
Discontinued usual care
(n = 0)
Analysed (n = 246)
Incomplete cases excluded
(n = 18)
110% recruitment
Intervention (n = 263)
Received intervention
training (n = 241)
80% completed
intervention
Did not attend follow up
(n=26)#
Discontinued intervention
(n = 53)
91% follow up
Incomplete cases excluded
(n = 29)
Analysed (n = 234)
Baseline Results
Results - primary outcome SBP
Results – secondary outcomes DBP
Results - subgroups
Results - medications

212 (80%) self managed for full 12 months

148 (70%) made at least one
medication change

At 12m intervention group prescribed
0.46 (0.34, 0.58) additional antiHT (p=0.001)

Main changes seen in thiazides and
calcium channel blockers
(60% on ACEI/ARB at baseline)
Results – side effects

Similar side effects in intervention vs control
Treatment targets

Observational data shows that achieved
blood pressure correlates with CVD outcome
 Wald meta analysis suggests that treatment
effects similar regardless of baseline
– But low baseline BP trials almost exclusively
secondary prevention
Treatment targets

Little convincing benefit from lower than
140/90 mmHg target in uncomplicated HT
 Evidence for systolic targets sparse
 Note reduced targets if out of office measure
What about old people?
Meta analysis – 6701 patients; mean FU 3.5 yrs; mean entry SBP 175
Target 150 mmHg systolic; Mean reduction SBP around 12mmHg
Journal of Hypertension 2010, 28:1366–1372
Bottom Line

Consider enhanced use of out of office
measurement, especially for diagnosis

Ambulatory monitoring for diagnosis is cost
effective due to better targeting of treatment

Home monitoring useful for ongoing management

Patients can do it too!

140/90 mmHg best evidence target unless
secondary prevention or over 80