Ofili EO, Cohen JD, St. Vrain JA, Pearson A, Martin TJ, Uy ND, Castello R, Labovitz AJ. Effect of Treatment of Isolated Systolic Hypertension on Left Ventricular Mass. JAMA. 1998;279(10):778-780. doi:10.1001/jama.279.10.778
From the Section of Cardiology, Department of Internal Medicine, Morehouse School of Medicine, Atlanta, Ga (Dr Ofili); Division of Cardiology, Department of Internal Medicine, St Louis University Medical Center, St Louis, Mo (Drs Cohen, Martin, Uy, Castello, and Labovitz and Ms St. Vrain); and Division of Cardiology, Ohio State University College of Medicine, Columbus (Dr Pearson).
Context.— Left ventricular (LV) hypertrophy is a common problem among elderly
patients with isolated systolic hypertension (ISH), but the effect of treatment
of ISH on LV mass is not known.
Objective.— To assess the ability of antihypertensive drug treatment to reduce LV
mass in ISH.
Design.— Echocardiographic Substudy of the Systolic Hypertension in the Elderly
Patients.— A total of 104 participants at the St Louis SHEP site who had interpretable
baseline echocardiograms, 94 of whom had 3-year follow-up echocardiograms.
Intervention.— The SHEP participants were randomized to placebo or active treatment
with chlorthalidone (12.5-25 mg/d), with atenolol (25-50 mg/d) added if necessary
to maintain goal blood pressure.
Main Outcome Measure.— Change in LV mass assessed by echocardiography.
Results.— Minimum follow-up was 3 years. In the active treatment group, 91% and
80% of subjects were receiving treatment with chlorthalidone alone by the
end of years 1 and 3, respectively. The LV mass index was 93 g/m2
in the active treatment group and 100 g/m2 in the placebo group
(P<.001). The LV mass index declined by 13% (95%
confidence interval, − 3% to − 23%) in the active treatment group
compared with a 6% increase (95% confidence interval, − 3% to + 16%)
in the placebo group over 3 years (P=.01).
Conclusion.— Treatment of ISH with a diuretic-based regimen reduces LV mass.
THE REPORT of the final results of the Systolic Hypertension in the
Elderly Program (SHEP) drew attention to the problem of isolated systolic
hypertension and conclusively showed the value of treatment.1
In this landmark study there was a 36% reduction in total stroke incidence,
the primary end point, among those randomly assigned to active antihypertensive
treatment compared with the placebo control group. In addition to the blood
pressure level, left ventricular hypertrophy (LVH) is a powerful and independent
predictor of cardiovascular morbidity and mortality.2- 6
Echocardiographically determined left ventricular (LV) mass is the most
powerful risk factor for cardiovascular disease, yielding prognostic information
beyond that provided by traditional cardiovascular risk factors, including
high blood pressure.2 We have reported the
high prevalence of LVH in the SHEP echocardiographic cohort compared with
age-matched normotensive controls,7 and preliminary
results for the effect of treatment on LV mass.8,9
We now report the long-term effects of treatment with antihypertensive drugs
on echocardiographic LV mass in this SHEP cohort.
The rationale, design, and eligibility criteria of SHEP have been previously
described.10 Participants were at least 60
years of age, with an average systolic blood pressure of 160 mm Hg or greater
and diastolic blood pressure lower than 90 mm Hg. Exclusion criteria were
systolic blood pressure greater than 220 mm Hg; recent myocardial infarction
or coronary artery bypass surgery; and history of stroke, renal insufficiency,
alcohol abuse, or other serious coexisting conditions. Eligible volunteers
were randomized in a double-blind manner to active antihypertensive drug treatment
or placebo. A stepped-care program was used to achieve a goal of systolic
blood pressure between 140 and 159 mm Hg, depending on the baseline average.
Chlorthalidone was the step 1 drug, used in 2 progressive dosages (of 12.5
and 25 mg/d) if necessary to achieve the goal blood pressure. The step 2 drug
(atenolol, 25-50 mg/d) was added if needed to achieve the goal blood pressure.
Reserpine, 0.05 mg/d, was used if atenolol was contraindicated because of
significant bradycardia, bronchospasm, or heart failure. Blood pressure was
monitored at 1- to 3-month intervals, and clinical assessments of the participants
were undertaken at annual visits, according to the SHEP protocol.1
Between March 1985 and September 1987, approximately 30000 subjects
60 years of age and older underwent screening for SHEP eligibility in the
St Louis metropolitan area. From this group, 397 were eventually entered into
the study and were randomly assigned to drug or placebo. Beginning in May
1986, SHEP volunteers were asked to participate in the echocardiographic substudy,
which was added to the usual SHEP protocol. Of the 104 individuals who agreed
and who had interpretable baseline echocardiograms, 94 were available at 3
years of follow-up. These 94 participants who form the basis of this report
were no different from the 4736 participants in the entire SHEP study with
regard to age, race and sex ratios, baseline blood pressure, cholesterol level,
history of diabetes, and cigarette smoking.
All echocardiographic studies were performed in standard, parasternal
long- and short-axis views using a commercially available, high-resolution,
phased-array system with a 2.25- or 3.5-MHz transducer and were recorded on
0.5-in videotape. Two-dimensional directed M-mode recordings of the left ventricle
were made according to the recommendations of the American Society of Echocardiography.11 Digitized M-mode tracings were measured off line
for an average of 5 cardiac cycles. Echocardiographic measurements were made
without knowledge of the subjects' randomization status. The LV mass was derived
according to American Society of Echocardiography standards (LV Mass = 0.80
[1.04 × (LVEDD + VS + PW)3−LVEDD3], where
LVEDD indicates LV end diastolic dimension; VS, ventricular septum thickness;
and PW, posterior wall thickness) and indexed to body surface area to determine
the LV mass index.
Data were analyzed by intention to treat. All values are reported as
mean ± SD unless otherwise indicated. Analysis of variance for repeated
measures was used to compare baseline and follow-up data in the active treatment
and placebo groups. Analysis of variance with interaction plots was used to
examine the effects of treatment assignment, baseline LV mass, age, sex, and
total cholesterol level on LV mass at follow-up. The significance of any interaction
was analyzed using the Scheffé F test. Multiple regression analysis
was used to determine the independent effect of systolic and diastolic blood
pressure on LV mass. P<.05 was considered statistically
Of the 94 participants available for follow-up at 3 years, 47 were in
the active treatment group and 47 were in the placebo control group. The baseline
characteristics of the active and placebo groups were similar (Table 1). By the end of years 1 and 3, respectively, 91% and 80%
of patients in the active treatment group were receiving only the first-line
drug, chlorthalidone; 18% were receiving atenolol in addition to chlorthalidone,
while 2 participants were receiving reserpine at the end of year 3. Twenty-one
participants were receiving additional antihypertensive therapy prescribed
by their private physicians: 9 patients (17%) in the active treatment group
vs 12 patients (28%) in the placebo group (P= .39).
The following nonstudy drugs were used: calcium channel blockers in 9 patients
(active, 5; placebo, 4), diuretics in 7 patients (active, 2; placebo, 5),
angiotensin-converting enzyme inhibitors in 3 patients (active, 1; placebo,
2), and β-blockers (active, 2). The average systolic and diastolic blood
pressures from baseline through 3 years of follow-up are shown in Figure 1.
Reductions in both systolic and diastolic blood pressures occurred early
in the active treatment group (within 6 months of the start of therapy) and
were maintained throughout the 36-month follow-up period. The systolic and
diastolic blood pressures were significantly lower in the active treatment
group than in the placebo group at follow-up (systolic: 144 ± 15 vs
163 ± 21 mm Hg, P<.001; diastolic: 62 ±
8 vs 75 ± 9 mm Hg, P<.001).
Echocardiographic parameters were similar in the active treatment and
placebo groups at baseline (Table 1).
At follow-up, the body surface area remained unchanged. However, the LV mass
index and ventricular septal thickness were significantly lower in the active
treatment group than in the placebo group (LV mass index: 93 ± 28 vs
110 ± 26 g/m2, P<.001; ventricular
septal thickness: 1.02 ± 0.20 vs 1.17 ± 0.28 cm, P=.01).
Within-group analysis showed that the LV mass at follow-up was significantly
lower in the active treatment group compared with baseline (93 ± 28
g at follow-up vs 104 ± 30 g,P=.04), while
LV mass in the placebo group increased over the 3-year follow-up period (110
± 26 g at follow-up vs 106 ± 24 g, P=.05).
The wall thickness also decreased in the active treatment group compared with
baseline: 1.02 ± 0.20 cm at follow-up vs 1.07 ± 0.23 cm at baseline
(P=.69) for the ventricular septum and 0.93 ±
0.21 vs 1.07 ± 0.28 cm (P=.03) for the posterior
wall, while the septal thickness increased in the placebo group: 1.17 ±
0.28 vs 0.99 ± 0.16 cm (P=.03). There was
a nonsignificant reduction in the LV end diastolic dimension in the active
treatment group compared with the placebo group at follow-up (4.72 ±
0.88 cm vs 4.94 ± 0.82 cm, P=.44).
The mean changes (with 95% confidence intervals [CIs]) in blood pressure
and LV mass are shown in Figure 2.
The active treatment group had a 24% reduction in systolic blood pressure
(95% CI, − 17% to − 31%) and a 17% reduction in diastolic blood
pressure (95% CI, − 14% to − 20%). For the placebo group, the
mean systolic blood pressure reduction was − 11% (95% CI, − 2%
to − 20%) and the mean diastolic blood pressure reduction was −
9% (95% CI, − 5% to − 14%). There was a 13% reduction in mass
in the active treatment group (95% CI, − 3% to − 23%), compared
with a 6% increase in LV mass in the placebo group (95% CI, − 3% to
+ 16%) (P=.01).
Change in systolic blood pressure was correlated with change in LV mass
(r=0.40, P<.003) by multivariate
analysis, independent of treatment assignment, age, sex, and high blood cholesterol
level. Change in diastolic blood pressure showed a weaker correlation with
change in LV mass (r=0.30, P<.04).
After controlling for change in blood pressure, there were no independent
effects of treatment assignment (active treatment vs placebo), age, or cholesterol
level on change in LV mass with treatment. Although men had higher LV mass
at baseline and follow-up, both men and women in the active treatment group
had lower LV mass at follow-up than those in the placebo group (men: 99 ±
30 vs 119 ± 23 g, P= .01; women: 86 ±
25 vs 98 ± 21 g, P=.09). Active treatment
participants receiving chlorthalidone alone at dosages of 25 and 12.5 mg/d
had follow-up LV mass index values of 87 ± 26 and 93 ± 31 g/m2, respectively (P=.06), significantly lower
than values in the placebo group (P<.001). Only
8 active treatment participants received atenolol, 50 or 25 mg/d, in addition
to chlorthalidone; their LV mass index values at follow-up were 71 ±
18 and 103 ± 14 g/m2, respectively (P=.67). Open-label treatment did not significantly affect LV mass in
either the active treatment or placebo group. The LV mass index value for
the 9 active treatment participants receiving open-label treatment was 89
± 29 g/m2, compared with 95 ± 28 g/m2
for active treatment participants who did not receive open-label treatment
(P=.75). The LV mass index for the 12 patients in
the placebo group who received open-label treatment was 103 ± 24 g/m2, compared with 112 ± 26 g/m2 for placebo participants
who did not receive open-label treatment (P=.36).
The antihypertensive treatment regimen, using chlorthalidone as sole
therapy in 80% of the active treatment group, was effective in significantly
lowering systolic blood pressure and decreasing LV mass compared with the
placebo group. In the placebo group, LV mass increased over time, presumably
reflecting the natural history of isolated systolic hypertension. The change
in LV mass was significantly correlated with change in both systolic and diastolic
blood pressures (r=0.40 and r=0.30,
respectively) on multiple regression analysis. The reduction in LV mass in
the active treatment group was due to an actual decrease in wall thickness;
the decrease in LV end diastolic dimension (5%, compared with 2% in the placebo
group), although small, was directionally compatible with an additional effect
of diuretic agents to reduce LV chamber size and volume. The LV fractional
shortening, an index of systolic function, remained unchanged despite significantly
decreased LV mass, a finding consistent with previous studies of LV mass regression
in treated hypertension, which showed preservation of systolic function despite
a reduction in LV mass.12- 15
The SHEP study demonstrated that reducing blood pressure among persons
with isolated systolic hypertension is associated with a reduction in both
morbidity and mortality associated with cardiovascular disease.1
One of the most striking beneficial effects of treatment was the relative
risk of 0.46 for LV failure in the treatment group compared with the control
group. To the extent that concentric LV hypertrophy is a risk factor for LV
failure, our data are consistent with this finding. The various agents used
to lower blood pressure among patients with hypertension may have differing
effects on LV mass.16- 18
The efficacy of a diuretic-based regimen on LV mass reduction is consistent
with the findings of 2 multicenter trials of hypertension treatment in younger
This randomized study clearly shows that chlorthalidone therapy is effective
in reducing LV mass in elderly patients with isolated systolic hypertension.