CARDIOVASCULAR JOURNAL OF AFRICA • Vol 21, No 5, September/October 2010
298
AFRICA
The J-shaped curve: fact or fiction?
Dr Peter Meredith, Division of Cardio-
vascular andMedical Sciences, University
of Glasgow, Scotland
Current conventional wisdom suggests
that when it comes to good blood pressure
control, the lower the better. However,
because a pressure of zero over zero
is incompatible with life, there would
appear to be a J-shaped curve and there-
fore there must be a point at which lower
is not necessarily better.
‘This implies that we need to identify
the nadir of that curve to ensure that we
don’t inadvertently increase risk when
lowering blood pressure by intervening
too intensively’, said Dr Meredith. He was
addressing a meeting at the South African
Heart Association’s annual conference at
Sun City in August.
Most blood pressure treatment is guid-
ed by epidemiological data that suggest
that for every 20-mmHg rise in systo-
lic blood pressure, there is a doubling of
cardiovascular mortality risk. Dr Meredith
feels that too much of current practice is
guided by data that might well be inap-
propriate. ‘We can’t be certain of the
therapeutic implications and it requires
a leap of faith to translate epidemiology
into treatment, especially in the case of
blood pressure that is not especially high
to begin with. Matters are complicated by
the fact that patients are not homogene-
ous. Heart failure patients, for example,
are not the same as hypertensives and in
the context of heart failure, higher blood
pressure is indicative of better cardiac
function. A cautious approach is therefore
needed in these patients, and all factors
need to be monitored.’
Dr Meredith cited a number of stud-
ies, but underscored their limitations.
‘We really need outcomes trials, of which
there are a relative paucity. As a result,
much of what we think we know is based
on retrospective interpretations of exist-
ing trials.
In a 1987 study, Cruickshank argued
that the J-shaped curve was more appar-
ent in ischaemic heart disease. In the
INVEST study, Messerli found that a
diastolic blood pressure lower than 70
mmHg was associated with an increased
risk of myocardial infarction, an interest-
ing but far-from-definitive finding.
Sleight’s analysis of the ONTARGET
trial showed that while higher blood pres-
sure was associated with a linearly higher
stroke risk, when it came to composite
outcomes (including myocardial infarc-
tion and cardiovascular mortality), there
was virtual equivalence between those
with the highest and lowest blood pres-
sures, seemingly supporting the idea of
the J-shaped curve.’
Dr Meredith was directly involved in a
similar
post-hoc
analysis of the ACTION
trial, the first ever randomised, placebo-
controlled clinical trial of an anti-anginal
drug, nifedipine XL. The study’s overall
finding was that blood pressure is an
important risk factor in angina patients
and that treatment with nifedipine was
beneficial.
While Dr Meredith’s findings in
respect of stroke echoed Sleight’s on
ONTARGET, this was not the case with
regard to the other endpoints where the
lowest blood pressure was consistently
associated with the smallest risk. ‘The
prognostic value of baseline systolic and
diastolic blood pressure was such that
lowest risk was equated with lowest
pressure, a different finding from both
INVEST and ONTARGET.’
Turning to cardiovascular outcomes
with more- versus less-intensive blood
pressure lowering, Dr Meredith stated that
there were relatively consistent findings
in favour of the former, especially in high-
risk patients. ‘Intensively treated diabetic
patients in the ACCORD study, for exam-
ple, suffered no deleterious effects and
all their outcomes showed a trend in a
favourable direction.’
He feels that, in the main, researchers
are too obsessed with
p
-values and that
statistical significance may not necessar-
ily translate into therapeutic benefit. ‘I am
firmly in the camp that believes lower to
be better. Nothing suggests that this is bad
for the patient, when therapy is tailored
to the individual needs of the patient in
question.’
He argued too that in the older place-
bo-controlled trials that seem to show
a J-shaped curve, the increased risk of
events was not related to antihyperten-
sive treatment, but rather to poor health
conditions at baseline. ‘Our evidence
from ACTION suggests that there is
no J-shaped curve. Neither is there a
nifedipine-induced curve. That said, I
acknowledge that there are limitations to
our findings, but the same is true for all
the other analyses.’
He concluded by advising that some
caution should be exercised in accom-
modating the individual needs of patients.
‘However, failure to control blood pres-
sure remains a major problem and, instead
of getting side-tracked and enmeshed
in debates about the J-shaped curve, we
need rather to focus our energies on
achieving good control.’
