Download A meta-analysis of randomised controlled trials (RCT

Journal of Human Hypertension (1999) 13, 367–374
 1999 Stockton Press. All rights reserved 0950-9240/99 $12.00
http://www.stockton-press.co.uk/jhh
ORIGINAL ARTICLE
A meta-analysis of randomised controlled
trials (RCT) among healthy normotensive
and essential hypertensive elderly patients
to determine the effect of high salt (NaCl)
diet on blood pressure
S Alam and AG Johnson
The University of Queensland Department of Medicine, Princess Alexandra Hospital, Brisbane,
Queensland, Australia
To examine the effect of chronic NaCl ingestion on
blood pressure (BP) in the elderly, a meta-analysis was
undertaken of 11 randomised controlled trials of which
five included patients ⭓60 years of age only and six
included patients with a mean age close to 60 years. The
following databases were used: Medline, Embase, Current Contents, The Cochrane Library, the AMI and IPA
databases. Mean erect systolic and diastolic blood
pressures (SBP/DBP) on chronic (⭓9 weeks) high and
low NaCl diets were recorded, the pooled mean effect,
the pooled standard error and 95% confidence intervals
(CI) were calculated and linear regression was used to
evaluate the potential association between NaCl intake
and BP. When all trials were pooled, a chronic high NaCl
diet significantly increased mean SBP and DBP by 5.58
mm Hg (95%CI 4.31–6.85) and 3.5 mm Hg (95%CI 2.62–
4.38) respectively. There was a significant association
between the level of NaCl intake and SBP (P ⴝ 0.05, r2
ⴝ 0.37) but not DBP (P ⴝ 0.76, r2 ⴝ 0.01). When trials
were pooled separately, a chronic high NaCl diet
increased SBP by 5.46 mm Hg (95%CI 3.56–7.36) and
DBP by 2.63 mm Hg (95%CI 1.18– 4.08) in trials including
patients ⭓60 years of age only, and increased SBP by
3.27 mm Hg (95%CI 1.23–5.31) and DBP by 2.69 mm Hg
(95%CI 1.44 –3.94) in trials including patients with a
mean age close to 60 years. These data suggest that a
chronic high NaCl diet in elderly patients with essential
hypertension is associated with an increase in SBP and
DBP, the association is significant for both SBP and
DBP but more marked for SBP than DBP, the effect is
more pronounced the older the patient and NaCl dose
strongly predicts SBP in older patients.
Keywords: meta-analysis; randomised trials; elderly; essential hypertensive; high NaCl; blood pressure
Introduction
Isolated systolic hypertension (ISH) defined as systolic blood pressure (SBP) ⭓160 mm Hg and diastolic blood pressure (DBP) ⬍90 mm Hg is prevalent
in the elderly with a substantial morbidity and mortality.1 In the Intersalt study, an age associated
increase in SBP was positively related to the mean
Na excretion of individuals in each country raising
the importance of chronic high Na intake for the
initiation and/or maintenance of hypertension.2 In
addition, the elderly have been reported to have a
higher prevalence of sensitivity of their BP
responses to ingested NaCl than younger individuals.3–5 Na retention and hypertension in the elderly
is possibly due to a complex interaction between
neuroendocrine factors and the kidney.6 Although
several meta-analysis have been reported in younger
Correspondence: Shahin Alam, Department of Medicine, University of Queensland, Princess Alexandra Hospital, Ipswich Road,
Woolloongabba, Brisbane, QLD 4102, Australia
Received 13 November 1998; revised 1 February 1999; accepted
2 February 1999
adults to show a modest hypertensive effect of
chronic high NaCl intake,7 there have been none
published which were focused on elderly individuals. We therefore undertook a meta-analysis of elderly subjects to evaluate the effect of chronic high
NaCl diet on their BP and to examine for differential
effects on SBP and DBP.
Materials and methods
Computerised English-language literature searches
were undertaken of published randomised human
trials evaluating the effect of changes in dietary NaCl
on BP in healthy, essential hypertensive individuals. These searches included Medline Express
(January 1966 to May 1998), Embase (1988–1998),
Current Contents (October 97–December 97 and January 98–April 98), AMI (1966–May 1998), API
(1970–May 98) and the Cochrane Library. Search
strategies included the following terms: ‘hypertens*’,
‘hypertension’, ‘blood pressure’, ‘isolated systolic
hypertension’, ‘salt’, ‘dietary salt’, ‘sodium diet’,
‘sodium chloride diet’, ‘salt diet’, ‘dietary sodium’,
‘Na diet’, ‘NaCl’, ‘elderly’, ‘age’, ‘aging’, ‘geriatr*’,
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
368
‘aged’, ‘meta-analysis’ and ‘randomised’, ‘random
allocation’, ‘random’, ‘control’, ‘clinical trials’,
‘meta-analysis’, ‘double-blind method’, ‘doubleblind study’, ‘double’ and ‘blind’. Bibiliographies of
review and primary articles were also explored.
The search yielded 251 papers. The methods and
results sections of all 251 selected manuscripts were
reviewed independently by two authors (SA & AJ).
Inclusion criteria were: randomised controlled trials
(RCT) of the effect of chronic NaCl ingestion on BP
in the elderly where BP and age data were provided.
The final decision to include or reject a study was
reached by consensus between two of the authors
(SA & AJ). It was our intention to score the quality
of each study as we planned to include only high
quality studies for this meta-analysis. However, we
found that by restricting studies for inclusion to
those with a randomised design the quality assessment score8 tended to be high for each trial. Consequently, no randomised trials were excluded based
on quality assessment score (average score ⬎70%).
However, only six trials involved elderly subjects
(⭓60 years). Although the two9,10 papers refer to different aspects they came from the same study and
represent similar data. However, for the meta-analysis one set of data can be used only. One9 of these
two trials was included in this meta-analysis
together with a further six trials involving subjects
with a mean age close to 60 years (ie, 59.5, 58.4,
58.6, 58.4 age range (50–64) and age ⬎50 years). The
remaining 239 trials were excluded due to the following reasons: five trials were randomised but did
not provide age data; 14 trials were randomised but
failed to provide BP data adequate for the metaanalysis; one trial measured BP changes with consumption of dietary NaCl but randomised the study
subjects into two groups on the basis of potassium:
magnesium salt with both patients and controls
receiving NaCl; one trial included patients ⬎60
years but did not provide BP data on dietary changes
of Na because the study intention was to measure
heart rate variability with alteration in Na diet; and
218 trials were randomised but studied younger
individuals. Consequently, relevant identifying data
were recorded (Table 1) for 11 trials9,11–20 which
were included in the meta-analysis.
Relevant data were extracted by two authors (SA &
AJ) independently, including year of publication,
population characteristics, diagnosis, intervention,
study duration and measurements (BP, heart rate,
weight, urinary Na). Discrepancies in data extraction
were noted by both authors and a subset was compared for differences in data extracted–no differences were identified (r = 1). Where BP data were
provided (only in graphic form) it was estimated
from the published figures available. Where a
quantification of NaCl intake was not published, 24h urinary Na was used as a surrogate.
The RCTs evaluating the effect of high NaCl vs
low NaCl on healthy essential hypertensive elderly
patients were pooled into two broad categories—
those trials where all participants were aged above
60 years (five trials) (Tables 3A, 3B) and those where
the mean age was close to 60 years (four trials)
(Tables 4A, 4B)—two trials were excluded. One19
because of a substantially greater difference in Na
dose and the second20 due to a substantially smaller
difference in NaCl dose from the remaining studies.
However, all trials included for analysis were randomised, evaluated BP effects of high vs low chronic
NaCl dose and studied healthy older patients with
essential hypertension. Therefore, it was considered
appropriate to pool data from all 11 studies.
The mean difference, standard errors and standard
deviation of SBP and DBP between high and low
NaCl diet for each trial was calculated.21,22 Ninetyfive percent confidence intervals (95%CI) for each
trial were also calculated for mean erect SBP and
DBP (Table 2A and B, respectively).23 Standard error
of the difference between means (s.e.) was calculated for each trial.22 There was insufficient data for
other parameters including heart rate and 24-h urinary Na to be included in the meta-analysis.
The pooled estimate of the standard deviation
(s.d.) and variance of the difference (Vd) between
the effect of high NaCl and low NaCl diets for each
trial was calculated.21 The pooled mean effect of
high NaCl compared with low NaCl diets on 24-h
average SBP and DBP was determined by using the
standardised mean effect weighted by the inverse of
the variance of the difference for each trial.24,25 The
pooled standard error and the corresponding
95%CIs were then calculated.24 Linear regression
modelling was used to evaluate the potential association between high dietary NaCl intake and BP. The
standardised weighted mean was calculated by
dividing the pooled mean treatment effect by the
associated standard error.26
Results
Table 1 shows a summary of the 11 trials which met
the study criteria and were included in the metaanalysis. Eight trials were crossover design with
three using a parallel design. Nine involved essential hypertensive patients and the remaining two12,13
used normotensive subjects. Because the mean BP
changes in normotensive and hypertensive subjects
were comparable, they were pooled. Of the total 485
subjects, 92.8% completed these trials. The average
trial sample size was 44 and the average trial quality
assessment score was 73. The 95%CI for each trial
was estimated for mean erect SBP and DBP (Table
2A, B, respectively). Nine of 11 trials provided
stratification by sex and overall male:female ratio
was 2.11:1. Eleven trials with 90.9% performed in
English speaking countries were pooled into three
different ways (Tables 3 to 5). SBP and DBP data
averages for each trial was pooled separately. For
36% of the trials, the diet was either not specified
or stated to be normal. Trial duration was ⭓9 weeks
for all studies included. The average trial duration
was 25.4 weeks and the treatment period was from
9 to 104 weeks.
Tables 3A and 3B included subjects aged ⭓60
years and weighted pooled mean increases in erect
SBP and DBP were 5.46 mm Hg (pooled estimates of
95%CIs were 3.56–7.36) and 2.63 mm Hg (CIs 1.18–
4.08) respectively, due to increased NaCl consumption (low to high NaCl diet). The standardised
Table 1 Study details of randomised, controlled trials (1978–1997)
Age
(years)
Sex
Study
design
Low Salt
diet
(mmol/d)
High Salt
diet
(mmol/d)
Sample
size
Drop-out
Study
duration
(weeks)
Quality
assessment
score
Ethic’s
approval
Site
Nestel et al,
199311
Palmer
et al, 198912
Fotherby
et al, 19979
Schorr et al,
199613
Cappuccio
et al, 199714
Weir et al,
199515
Lancet, 198916
60–79
36 M
30 F
1M
6W
4M
13 W
10 M
11 W
24 M
23 W
14 M
8F
93 M
18 F
7M
Crossover
80
160
70
4
10
77
Yes
Australia
Crossover
43
175
7
1
16
67
Yes
USA
Crossover
80
160
17
0
14
80
Unknown
UK
Crossover
⭐100
⭐360
21
5
18
83
Yes
Germany
Crossover
80
200
52
5
10
86
Yes
UK
Crossover
40
200
22
0
9
83
Yes
USA
Parallel
80
160
111
8
14
67
Yes
Australia
Crossover
10
250
7
2
10
59
Unknown
USA
Crossover
80
160
88
0
22
70
Yes
Australia
Parallel
100
138
28
3
52
71
Unknown
UK
62
0
104
58
Unknown
Australia
Kurtz et al,
198719
Clin Ex Hyp,
198917
Silman et al,
198318
Morgan et al,
197820
78–96
66–79
60–72
60–78
59.5 ± 3*
58.4 ± 1*
58.4
58.6 ± 1.1*
50–64
⬎50
73 M
15 F
28
(not reported)
62
(not reported)
Parallel
157 (urinary
Na)
180 (urinary
Na)
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
Study
reference
* Mean ± s.e.m. (standard error of the mean).
369
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
370
Table 2 Meta-analysis of randomised, placebo-controlled trials (1978–1997)
Trials
Nestel et al, 1993
Palmer et al, 1989
Fotherby et al, 1997
Schorr et al, 1996
Cappuccio et al, 1997
Weir et al, 1995
Lancet, 1989
Kurtz et al, 1987
Clin Exp Hyp, 1989
Silman et al, 1983
Morgan et al, 1978
Reference
11
12
9
13
14
15
16
19
17
18
20
Weighted pooled
mean change*
95% confidence
intervals
Weighted pooled
mean change†
A Systolic blood pressure
4.82
1.55–8.09
10.29
1.74 –18.84
6
(−0.62)–12.62
7.15
1.05–13.25
3.9
(−1)–8.8
5
(−1.35)–11.35
3.1
(−0.72)–6.92
16
10.45–21.55
3.6
(−0.56)–7.76
0.5
(−7.0)–8.0
7.6
2.6–12.6
95% Confidence
intervals
B Diastolic blood pressure
2.28
(−0.5)–5.06
5.43
(−0.98)–11.84
3
(−1.82)–7.82
2.86
(−1.61)–7.33
2.1
(−1.17)–5.37
1
(−3.39)–5.39
3.2
0.73–5.67
8
3.2–12.8
2.1
(−0.13)– 4.33
5.6
0.68–10.52
9.3
5.91–12.69
Weighted pooled mean change of *SBP and †DBP with high salt diet.
Table 3A Pooled mean SBP with low and high salt diet
Study
details
Nestel et al, 1993
Palmer et al, 1989
Fotherby et al, 1997
Schorr et al, 1996
Cappuccio et al, 1997
Age (years)
60–79
78–96
66–79
60–72
60–78
Low Na SBP
(erect)
(mm Hg)
122.91 ± 9.55*
144.14 ± 8.77#
168 ± 25*
125.71 ± 15.71*
151.1 ± 21.2*
High Na DBP
(erect)
(mm Hg)
127.73 ± 13.18*
154.43 ± 10.22#
174 ± 22*
132.86 ± 22.86*
155 ± 21.5*
Weighted
pooled
mean
change
(mm Hg)
s.d.
difference
Variance
difference
(Vd)
4.82
10.29
6
7.15
3.9
11.51
23.32
23.55
19.61
21.35
2
13.46
8.08
6.93
4.4
* Mean ± s.d. (standard deviation).
# Mean ± s.e.m. (standard error of the mean).
Pooled mean treatment effect: 5.46 mm Hg.
Associated standard error: 0.97.
Pooled estimates of 95% confidence intervals: 3.56–7.36.
Table 3B Pooled mean DBP with low and high salt diet
Study
details
Nestel et al, 1993
Palmer et al, 1989
Fotherby et al, 1997
Schorr et al, 1996
Cappuccio et al, 1997
Age (years)
Low Na DBP
(erect)
(mm Hg)
High Na DBP
(erect)
(mm Hg)
Weighted
pooled
mean
change
(mm Hg)
s.d.
difference
Variance
difference
(Vd)
60–79
78–96
66–79
60–72
60–78
72.45 ± 7.36*
70.86 ± 4.85#
99 ± 12*
70 ± 9.29*
91.6 ± 8.7*
74.73 ± 9*
76.29 ± 5.83#
102 ± 13*
72.86 ± 11.43*
93.7 ± 10.5*
2.28
5.43
3
2.86
2.1
8.22
13.13
12.51
10.42
9.64
1.43
7.58
4.29
3.86
1.99
* Mean ± s.d. (standard deviation).
# Mean ± s.e.m. (standard error of the mean).
Pooled mean treatment effect: 2.63 mm Hg.
Associated standard error: 0.74.
Pooled estimates of 95% confidence intervals: 1.18– 4.08.
weighted mean26 increase in SBP was 1.59 times
more when compared with DBP, which suggests that
the effect of salt on SBP is greater than on DBP.
Tables 4A and 4B included four and excluded two
trials because of data inconsistency. From the
remaining four trials, subjects with a mean age
below but close to 60 years and weighted pooled
mean increase in erect SBP and DBP of 3.27 mm Hg
(pooled estimates of 95%CIs were 1.23–5.31) and
2.69 mm Hg (CIs 1.44 –3.94) respectively due to
increased NaCl consumption (low to high NaCl
diet). The standardised weighted mean26 increase in
SBP was 0.75 times more when compared with DBP.
Tables 5A and 5B included 11 trials (patients with
age ⬎60 years and age close to 60 years). When combined, the weighted pooled mean increase in erect
SBP and DBP was 5.58 mm Hg (pooled estimates of
95% CIs were 4.31–6.85) and 3.5 mm Hg (CIs 2.62–
4.38) respectively, due to increased NaCl consumption (low to high NaCl diet) (Tables 5A and 5B,
respectively). The standardised weighted mean
increase in SBP was 1.10 times more when com-
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
Table 4A
371
Pooled mean SBP with low and high salt diet
Study
details
Weir et al, 1995
Lancet, 1989
Clin Exp Hyp, 1989
Silman et al, 1978
Age (years)
59.5 ± 3*
58.4 ± 1*
58.6 ± 1.1*
50–64 (ranged)
Low Na SBP
(erect)
(mm Hg)
157 ± 3*
149.1 ± 1.9*
152 ± 2.2*
138.6 ± 8.4*
High Na SBP
(erect)
(mm Hg)
162 ± 5*
152.2 ± 1.9*
155.6 ± 2.3*
139 ± 6.2*
Weighted
pooled
mean
change
(mm Hg)
s.d.
difference
Variance
difference
(Vd)
5
3.1
3.6
0.5
26.65
13.64
21.12
25.04
7.43
2.69
3.18
10.22
* Mean ± s.e.m. (standard error of the mean).
Pooled mean treatment effect: 3.27 mm Hg.
Associated standard error: 1.04.
Pooled estimates of 95% confidence intervals: 1.23–5.31.
Table 4B Pooled mean DBP with low and high salt diet
Study
details
Weir et al, 1995
Lancet, 1989
Clin Exp Hyp, 1989
Silman et al, 1978
Age (years)
Low Na DBP
(erect)
(mm Hg)
High Na DBP
(erect)
(mm Hg)
Weighted
pooled
mean
change
(mm Hg)
s.d.
difference
Variance
difference
(Vd)
59.5 ± 3
58.4 ± 1
58.6 ± 1.1
50–64 (ranged)
95.5 ± 2.5*
91.4 ± 0.7*
95 ± 0.7*
80.9 ± 3.6*
96.5 ± 2.5*
94.6 ± 0.9*
97.1 ± 0.6*
86.5 ± 2.7*
1
3.2
2.1
5.6
11.73
5.75
6.12
10.83
3.54
1.13
0.92
4.42
* Mean ± s.e.m. (standard error of the mean).
Pooled mean treatment effect: 2.69 mm Hg.
Associated standard error: 0.64.
Pooled estimates of 95% confidence intervals: 1.44 –3.94.
pared with DBP, which suggests that the effect of
salt on SBP is greater than on DBP.
In 11 trials, the erect mean differences in SBP
were plotted against the dietary NaCl intakes (⌬, difference from low to high NaCl diets) achieved after
high Na diet (1978–1997) (Figure 1). The regression
analysis revealed a significant (P = 0.05) association
between the level of SBP and the level of NaCl
intake with Na intake accounting for 37% of the
variability in SBP participants of these trials.
In 11 trials, the erect mean differences in DBP
were plotted against the dietary change in NaCl
intake (⌬, difference from low to high NaCl diets)
achieved after high Na diet (1978–1997) (Figure 2).
The regression analysis showed that there was nonsignificant (P = 0.76) association between the level
of DBP and the level of NaCl intake.
Discussion
Two hundred and eighteen (218) of the 251 trials
initially found from different databases were
excluded because those trials studied either middleaged or younger groups with extensive differences
in age range. Of the remaining 33 trials, 21 matched
the study criteria but had to be excluded because of
inadequate or insufficient information or data
required for this meta-analysis. Finally, from the
remaining 12 trials, 11 were selected for the metaanalysis. One of two trials was included because
both provided similar data in two different journals
in different years. In the selected 11 trials, one of
the trials did not provide an age but age range (50–
64 years). This was included because all the subjects
were between 50 and 60 years and it was assumed
that the mean age may be close to 60. Another trial
was also included despite no given age or age range
because the trial mentioned that all the study subjects were above 50 years and it was assumed that
their age might be close to 60 years. Selection bias
was minimised by independent evaluation of
methods and results by two authors (SA & AJ).
When trials were pooled separately, ie, subjects
age ⭓60 years and mean age below but close to 60
years (which included most of the subjects close to
60 and also some middle-aged subjects) it was found
that chronic high NaCl intake increased both SBP
and DBP but the increment for SBP was much
greater than for DBP. When compared between
groups, chronic high NaCl intake caused a greater
increase in SBP in the former group (5.46 vs 3.27
mm Hg) but not for DBP (2.69 vs 2.63 mm Hg). The
above data suggested that SBP increased more than
DBP in both age groups studied but increases in SBP
in subjects whose age was ⬎60 years were more
when compared with increases in SBP in subjects
whose mean age was below but close to 60 years.
Results suggested a positive association with SBP,
high NaCl intake and aging. Consequently, dietary
restriction of NaCl might be useful in minimising
the elevation of SBP in older individuals.
When all trials were pooled, the meta-analysis
demonstrated that switching diets from low to high
NaCl increased SBP by 5.58 mm Hg and DBP by 3.5
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
372
Table 5A Pooled mean SBP with low and high salt diet (1978–1997)
Study details
Nestel et al, 1993
Palmer et al, 1989
Fotherby et al, 1997
Schorr et al, 1996
Cappuccio et al, 1997
Weir et al, 1995
Lancet, 1989
Kurtz et al, 1987
Clin Exp Hyp, 1989
Silman et al, 1978
Morgan et al, 1978
Age
(years)
60–79
78–96
66–79
60–72
60–78
59.5 ± 3
58.4 ± 1
58.4
58.6 ± 1.1
50–64 (ranged)
⬎50
Low Na SBP
(erect)
(mm Hg)
122.91 ± 9.55
144.14 ± 8.77
168 ± 25
125.71 ± 15.71
151.1 ± 21.2
157 ± 3
149.1 ± 1.9
126 ± 4
152 ± 2.2
138.6 ± 8.4
152.4 ± 2.5
High Na SBP
(erect)
(mm Hg)
127.73 ± 13.18
154.43 ± 10.22
174 ± 22
132.86 ± 22.86
155 ± 21.5
162 ± 5
152.2 ± 1.9
142 ± 4
155.6 ± 2.3
139 ± 6.2
160 ± 4
Weighted
pooled mean
change (mm Hg)
s.d.
difference
Variance
difference
(Vd)
4.82
10.29
6
7.15
3.9
5
3.1
16
3.6
0.5
7.6
11.51
23.32
23.55
19.61
21.35
26.65
13.64
8.94
21.12
25.04
18.57
2
13.46
8.08
6.93
4.4
7.43
2.69
5.65
3.18
10.22
4.72
Weighted
pooled mean
change (mm Hg)
s.d.
difference
Variance
difference
(Vd)
2.28
5.43
3
2.86
2.1
1
3.2
8
2.1
5.6
9.3
8.22
13.13
12.51
10.42
9.64
11.73
5.75
7.07
6.12
10.83
8.81
1.43
7.58
4.29
3.86
1.99
3.54
1.13
4.47
0.92
4.42
2.24
Pooled mean treatment effect: 5.58 mm Hg.
Associated standard error: 0.65.
Pooled estimates of 95% confidence intervals: 4.31–6.85.
Table 5B Pooled mean DBP with low and high salt diet (1978–1997)
Study details
Nestel et al, 1993
Palmer et al, 1989
Fotherby et al, 1997
Schorr et al, 1996
Cappuccio et al, 1997
Weir et al, 1995
Lancet, 1989
Kurtz et al, 1987
Clin Exp Hyp, 1989
Silman et al, 1978
Morgan et al, 1978
Age
(years)
60–79
78–96
66–79
60–72
60–78
59.5 ± 3
58.4 ± 1
58.4
58.6 ± 1.1
50–64 (ranged)
⬎50
Low Na DBP
(erect)
(mm Hg)
72.45
70.86
99
70
91.6
95.5
91.4
76
95
80.9
100.3
±
±
±
±
±
±
±
±
±
±
±
7.36
4.85
12
9.29
8.7
2.5
0.7
2
0.7
3.6
1
High Na DBP
(erect)
(mm Hg)
74.73
76.29
102
72.86
93.7
96.5
94.6
84
97.1
86.5
109.6
±
±
±
±
±
±
±
±
±
±
±
9
5.83
13
11.43
10.5
2.5
0.9
4
0.6
2.7
2
Pooled mean treatment effect: 3.5 mm Hg.
Associated standard error: 0.45.
Pooled estimates of 95% confidence intervals: 2.62– 4.38.
Figure 1 Erect mean change in SBP are plotted against the dietary
sodium intakes (⌬, difference from low to high salt diet) achieved
after high salt diet, in 11 trials (1978–1997). The regression analysis showed that there was a positive association between the level
of SBP and the level of salt intake (P = 0.05, R = 0.06). The results
also suggested that approximately 37% of variation an increase
in SBP was attributed to the change in salt intake (r2 = 0.37).
Figure 2 Erect mean change in DBP are plotted against the dietary
change in sodium intakes (⌬, difference from low to high salt diet)
achieved after high salt diet, in 11 trials (1978–1997). The
regression analysis showed that there was nonsignificant association between the level of DBP and the level of salt intake (P =
0.76, r2 = 0.01 and R = 0.10).
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
mm Hg. This suggests that in older individuals
increasing NaCl intake produce a significant elevation in BP, which is greater for SBP than for DBP.
When the increase in erect SBP data from all trials
were pooled with the dietary changes in NaCl
intake, linear regression analysis revealed a significant association between SBP but not DBP, and the
level of NaCl intake. This indicated that the level
of SBP in the elderly was positively and strongly
influenced by the level of NaCl consumption.
A reduction of 100 mmol Na excretion in hypertensive subjects from 25–55 years caused an average
9 mm Hg drop in SBP and represents a positive and
statistically significant (P ⬍ 0.001) association
between the slope of BP and age with Na excretion.
The association of NaCl intake and SBP for the older
age group compared with other age groups was
greater and more significant.27 In essential hypertensive subjects (25–80 years) high NaCl intake significantly elevated BP (P ⬍ 0.001) and pronounced saltsensitivity (SS, 91%) more in the older age group
only, suggesting that age predicts the BP changes
resulting from high NaCl intake and SS associated
with the age-related changes in renal and endocrine
function.28 Excessive NaCl intake may contribute to
the pathogenesis of hypertension, particularly SS
hypertension.29
How clinically important is reduction in SBP of
5.58 mm Hg and DBP of 3.5 mm Hg by reduced dietary NaCl intake? Law et al30 demonstrated that for
subjects with normal and high BP, a 50 mmol/day
reduction in NaCl intake for a few weeks can reduce
SBP by 5 and 7 mm Hg respectively, DBP by 2.5 and
3.5 mm Hg respectively and lower the occurrence of
ischaemic heart disease (IHD) by 15% and stroke by
26%. In an entire western population, such a NaCl
reduction would reduce BP by 5 mm Hg and prevent
mortality from IHD and stroke by 16% and 22%
respectively. By combining restricted NaCl of
50 mmol and 100 mmol/day with BP medication,
mortality from IHD and stroke would reduce by 20%
and 30% respectively.30,31
In the elderly population, the occurrence of
hypertension, particularly ISH, is rapidly increasing
and is an important determinant of cardiovascular
risk.32 Mechanisms possibly responsible for
developing BP in elderly subjects include reduced
arterial compliance, elevated peripheral vascular
resistance, alteration in cardiac output, reduction in
plasma renin activity and decreased beta-adrenergic
function. Additionally, environmental factors such
as diet, SS status and genetic predisposition may
also contribute to hypertension.33 Observational
studies within western populations revealed a
strong and positive association (P ⬍ 0.001) between
NaCl intake and BP.30,31 Our meta-analysis in the
elderly demonstrated that increased salt ingestion
caused a significant increase in SBP, suggesting a
possible pathogenetic role for salt in the initiation
and/or maintenance of ISH.
In conclusion, a chronic high salt dose in elderly
individuals with essential hypertension significantly elevated SBP and DBP, but the effect was
more marked for SBP when compared with DBP and
the older the patient. Therefore, chronic high salt
dose is an important determinant of SBP in elderly
hypertensive patients.
References
1 Lichtenstein MJ. Isolated systolic hypertension: how
common? How risky? South Med J 1985; 78: 972–978.
2 Nielsen WB, Svendsen TL, Rasmussen SP. Isolated
systolic hypertension in the elderly. Ugeskr Laeger
1994; 156: 1455–1459.
3 Nielsen WB, Vestbo J, Jensen GB. Isolated systolic
hypertension as a major risk factor for stroke and myocardial infarction and an unexploited source of cardiovascular prevention: a prospective population-based
study. J Hum Hypertens 1995; 9: 175–180.
4 Luft FC et al. Effects of ageon renal sodium homeostasis and its relevance to sodium Sensitivity. Am J
Med 1987; 82 (Suppl 1B): 9–15.
5 Grzegarczyk PB, Jones CW, Mistretta CM. Age related
differences in salt taste acuity. J Gerontol 1979; 324:
834 –860.
6 Campese VM. Salt sensitivity in hypertension. Renal
and cardiovascular implications (clinical conference).
Hypertension 1994; 23: 531–550.
7 Ogihara T et al. Guidelines on treatment of hypertension in the elderly 1995—a tentative plan for guidelines on treatment of hypertension in the elderly, comprehensive research projects on aging and health, the
Ministry of Health and Welfare of Japan. Nippon
Ronen Igakkai Zasshi 1996; 33: 945–975.
8 Chalmers TC et al. A method for assessing the quality
of a randomized control trial. Controlled Clin Trials
1981; 2: 31– 49.
9 Fotherby MD, Potter JF. Metabolic and orthostatic
blood pressure responses to a low sodium diet in elderly hypertensives. J Hum Hypertens 1997; 11: 361–
366.
10 Fotherby MD, Potter JF. Effects of moderate sodium
restriction of clinic and twenty-four-hour ambulatory
blood pressure in elderly hypertensive subjects. J Hypertens 1993; 11: 657–663.
11 Nestel PJ et al. Enhanced blood pressure response to
dietary salt in elderly women, especially those with
small waist:hip ratio. J Hypertension 1993; 11: 1387–
1394.
12 Palmer RM, Osterweil D, Lustig GL, Stern N. The effect
of dietary salt ingestion on blood pressure of old-old
subjects. A double-blind, placebo-controlled, crossover trial. J Am Geriatr Soc 1989; 37: 931–936.
13 Schorr U, Distler A, Sharma AM. Effect of sodium
chloride and sodium bicarbonate rich mineral water
on blood pressure and metabolic parameters in elderly
normotensive individuals: A randomized double-blind
crossover trial. J Hypertens 1996; 14: 131–135.
14 Cappuccio FP et al. Double-blind randomised trial of
modest salt restriction in older people. Lancet 1997;
350 (908): 850–854.
15 Weir MR, Dengel DR, Behrens MT, Goldberg AP. Saltinduced increases in systolic blood pressure affect
renal hemodynamics and proteinuria. Hypertension
1995; 25: 1339–1344.
16 Australian National Health and Medical Research
Council Dietary Salt Study Management Committee.
Fall in blood pressure with modest reduction in dietary salt intake in mild hypertension. Lancet 1989; 1
(8635): 399– 402.
17 Australian National Health and Medical Research
Council Dietary Salt Study Management Committee.
Effects of replacing sodium intake in subjects on a low
sodium diet: a crossover study. Clin and Exp Hypertens A 1989; 11: 1011–1024.
373
Meta-analysis of NaCl and BP in the elderly
S Alam and AG Johnson
374
18 Silman AJ, Mitchell P, Locke C, Humpherson P. Evaluation of the effectiveness of a low sodium diet in the
treatment of mild to moderate hypertension. Lancet
1983; 1 (8335): 1179–1182.
19 Kurtz TW, Al-Bander HA, Morris RC. Salt-sensitive
essential hypertension in men. Is the sodium ion alone
important? N Engl J Med 1987; 317: 1043–1048.
20 Morgan T et al. Hypertension treated by salt restriction. Lancet 1978; 1 (8058): 227–230.
21 Gardner MJ, Altman DG. Confidence intervals rather
than p values: Estimation rather than hypothesis testing. Br Med J Clin Res Ed 1986; 292 (6522): 746–750.
22 Kirkwood BR. Comparison of two means. In: Essential
of Medical Statistics. Blackwell Scientific Publications: London, 1988, pp 41– 45.
23 Kirkwood BR. Confidence interval for a mean. In:
Essential of Medical Statistics. Blackwell Scientific
Publications: London, 1988, pp 27–32.
24 Cappuccio FP, MacGregor GA. Does potassium supplementation lower blood pressure? A meta-analysis of
published trials. J Hypertens 1991; 9: 465–73.
25 Rosenthal R, Rubin DB. Comparing effect sizes of independent studies. Psychol Bull 1982; 92: 500–504.
26 Kirkwood BR. Measures of mortality and morbidity.
In: Essential of Medical Statistics. Blackwell Scientific
Publications: London, 1988, pp 106–117.
27 Elliott P. Observational studies on salt and blood
pressure. Hypertension 1991; 7: 13–18.
28 Overlack A et al. Age is a major determinant of the
divergent blood pressure responses to varing salt
intake in essential hypertension. Am J Hypertens 1995;
8: 829–836.
29 Carrageta MO, Negrao L, de-Padua F. Community
based stroke prevention: a Portuguese challenge.
Health Res 1994; 6: 189–195.
30 Law MR, Frost CD, Wald NJ. 111 Analysis of observational data within populations. BMJ 1991; 302:
815–818.
31 Law MR, Frost CD, Wald NJ. 111 Analysis of observational data within populations. BMJ 1991; 302:
819–824.
32 Sowers JR. Hypertension in the elderly. Am J Med
1987; 82 (Suppl 1B): 1–8.
33 Furmaga EM, Murphy CM, Carter BL. Isolated systolic
hypertension in older patients. Clin Pharm 1993; 12:
347–358.