You are on page 1of 7

Clinical Review & Education

JAMA Internal Medicine | Review

Reduced Salt Intake for Heart Failure


A Systematic Review
Kamal R. Mahtani, PhD; Carl Heneghan, DPhil; Igho Onakpoya, DPhil; Stephanie Tierney, MA, PhD;
Jeffrey K. Aronson, DPhil; Nia Roberts, MSc; F. D. Richard Hobbs, FMedSci; David Nunan, MSc, PhD

Invited Commentary
IMPORTANCE Recent estimates suggest that more than 26 million people worldwide have Supplemental content
heart failure. The syndrome is associated with major symptoms, significantly increased
mortality, and extensive use of health care. Evidence-based treatments influence all these
outcomes in a proportion of patients with heart failure. Current management also often
includes advice to reduce dietary salt intake, although the benefits are uncertain.

OBJECTIVE To systematically review randomized clinical trials of reduced dietary salt in adult
inpatients or outpatients with heart failure.

EVIDENCE REVIEW Several bibliographic databases were systematically searched, including


the Cochrane Central Register of Controlled Trials, MEDLINE, Embase, and CINAHL. The
methodologic quality of the studies was evaluated, and data associated with primary
outcomes of interest (cardiovascular-associated mortality, all-cause mortality, and adverse
events, such as stroke and myocardial infarction) and secondary outcomes (hospitalization,
length of inpatient stay, change in New York Heart Association [NYHA] functional class,
adherence to dietary low-salt intake, and changes in blood pressure) were extracted.

FINDINGS Of 2655 retrieved references, 9 studies involving 479 unique participants were
included in the analysis. None of the studies included more than 100 participants. The risks
of bias in the 9 studies were variable. None of the included studies provided sufficient
data on the primary outcomes of interest. For the secondary outcomes of interest,
2 outpatient-based studies reported that NYHA functional class was not improved by
restriction of salt intake, whereas 2 studies reported significant improvements in NYHA
functional class.

CONCLUSIONS AND RELEVANCE Limited evidence of clinical improvement was available


among outpatients who reduced dietary salt intake, and evidence was inconclusive for
inpatients. Overall, a paucity of robust high-quality evidence to support or refute current
guidance was available. This review suggests that well-designed, adequately powered
studies are needed to reduce uncertainty about the use of this intervention.

PROTOCOL REGISTRATION PROSPERO Identifier: CRD42015019504

Author Affiliations: Centre for


Evidence Based Medicine, Nuffield
Department of Primary Care Health
Sciences, University of Oxford,
Oxford, United Kingdom (Mahtani,
Heneghan, Onakpoya, Tierney,
Aronson, Hobbs, Nunan); Knowledge
Centre, Bodleian Libraries, Old Road
Campus, University of Oxford,
Oxford, United Kingdom (Roberts).
Corresponding Author: Kamal Ram
Mahtani, PhD, Centre for Evidence
Based Medicine, Nuffield Department
of Primary Care Health Sciences,
University of Oxford,
Radcliffe Observatory Quarter,
Woodstock Road,
JAMA Intern Med. doi:10.1001/jamainternmed.2018.4673 Oxford OX2 6GG, United Kingdom
Published online November 5, 2018. (kamal.mahtani@phc.ox.ac.uk).

(Reprinted) E1
© 2018 American Medical Association. All rights reserved.

Downloaded From: on 11/05/2018


Clinical Review & Education Review Reduced Salt Intake for Heart Failure

D
espite advances in diagnosis and management, the preva-
lence of heart failure is increasing.1-3 Recent estimates Key Points
suggest that more than 26 million people are affected Question What is the evidence that reduced dietary salt intake
worldwide, including inpatients and outpatients.4,5 benefits adults with heart failure?
Many high-quality trials and systematic reviews have shown that
Findings In this systematic review of 9 unique studies of 479
selected therapeutic interventions, including angiotensin-
unique patients with heart failure, an overall paucity of robust
converting enzyme inhibitors, angiotensin receptor antagonists, high-quality evidence was available to support or refute the use
β-adrenoceptor antagonists, and aldosterone receptor antago- of reduced dietary salt intake.
nists, produce significant and clinically relevant improvements in
Meaning Despite broad advocacy, uncertainty about the
many adverse outcomes in heart failure, at least in most patients with
robustness of advice to reduce salt intake in patients with
reduced ventricular function. Advice to reduce salt intake is recom- heart failure remains.
mended in several international cardiovascular guidelines6 (eTable
1 in the Supplement).
However, restriction of salt intake has uncertain efficacy.7 Ob- Data Extraction and Synthesis
servational studies have shown potential benefits and harms, whereas Data were independently extracted into a prespecified data extraction
some randomized clinical trials (RCTs) have suggested harms.8 A table. Study quality was assessed using the Cochrane risk of bias tool.14
previous systematic review and meta-analysis9 concluded that a low- The primary outcomes of interest were cardiovascular-
salt diet, compared with normal salt intake, significantly increased associated mortality, all-cause mortality, and adverse events (eg,
morbidity and mortality in individuals with heart failure. That review stroke, myocardial infarction, hyponatremia, and hypernatremia).
included 6 RCTs, all from a single research group in Italy. The review The secondary outcomes included hospitalization, length of inpa-
led to initial calls for a reevaluation of guidelines and current tient stay, change in clinical signs and symptoms of heart failure (eg,
practice.10,11 However, the systematic review was retracted after New York Heart Association [NYHA] functional class), measure-
concerns were expressed about the integrity of the data.12 ment and adherence to low dietary salt intake, and change in
The lack of high-quality evidence in this area is reflected in a joint blood pressure.
executive summary from the US National Heart, Lung, and Blood
Institute and the National Institutes of Health Office of Dietary
Supplements,13 who have called for better-quality evidence. A
Results
systematic review of the role of reduced dietary salt in individuals
with heart failure is therefore needed. Results of the Search
Our search results are shown in Figure 1. We retrieved 2655 refer-
ences after removing duplicates. At least 2 authors independently
screened these references and identified 27 studies of potential rel-
Methods evance. We reviewed those full texts and excluded 18, after search-
We conducted a systematic review following the guidance in the ing the references for other relevant studies (eTable 2 in the Supple-
Cochrane Handbook for Systematic Reviews of Interventions14 and ment). We therefore included 9 unique studies involving 479 unique
planned to perform a meta-analysis if sufficient reliable data were participants.15-23 We also identified 5 potentially relevant current
available. Further details are provided in the eMethods in the trials (eTable 3 in the Supplement). During our analysis, prelimi-
Supplement, including any differences between the protocol and the nary data from one of these trials was published in the form of an
final review. abstract.22 We included data from this abstract in our review and
corresponded with the lead author (Scott L. Hummel, MD, email
Data Sources and Searches communication, April 30, 2018).24
We included RCTs that evaluated the effects of reduced dietary salt Table 1 contains a summary of the 9 included studies, none of
intake in adults (aged ⱖ18 years) with heart failure. We systemati- which involved more than 100 participants. Eight studies were pub-
cally searched across several bibliographic databases, including the lished in English15-20,22,23 and 1 was published in Portuguese.21 Three
Cochrane Central Register of Controlled Trials, MEDLINE, Embase, studies were conducted in Europe,16,19,20 2 in South America,15,21 and
CINAHL, and ClinicalTrials.gov from inception to February 9, 2016. 4 in North America. 17,18,22,23 Two studies were conducted in
We continued to survey the literature until March 2, 2018, includ- inpatients15,21 and 7 were conducted in outpatients.16-20,22,23 Eight
ing an updated search of the Cochrane Central Register of studies were published as full texts15-21 and 1 in abstract form,22
Controlled Trials and MEDLINE. We also searched for publications although we supplemented the data from this abstract with
arising from current trials identified in ClinicalTrials.gov. additional information from the principal investigator and the
published protocol.24
Study Selection
We included studies that recruited outpatients with chronic stable Risk of Bias Assessment
heart failure and inpatients who had been admitted to hospital with The risks of bias in the 9 included studies were variable. None had
acutely decompensated heart failure. At least 2 authors indepen- a low risk of bias in each of 7 domains, and all were assessed as
dently screened and selected studies for potential inclusion and com- having an unclear risk of bias in at least 2 domains. The results are
pared their selections; disagreements were resolved through summarized in Figure 2. Further details of the assessment are
discussion with the remaining authors. provided in eResults 1 in the Supplement.

E2 JAMA Internal Medicine Published online November 5, 2018 (Reprinted) jamainternalmedicine.com

© 2018 American Medical Association. All rights reserved.

Downloaded From: on 11/05/2018


Reduced Salt Intake for Heart Failure Review Clinical Review & Education

Description of Interventions Figure 1. Study Flow Diagram


We used the Template for Intervention Description and Replication
(TIDieR) tool25 to improve the completeness of our reporting and 2679 Records identified through
the potential replicability of the interventions we identified (eTable database searching
4 in the Supplement). We found notable clinical and methodologic
heterogeneity across the studies. Several studies did not provide 2655 Records after duplicates removed
adequate descriptions of the intervention, how the intervention
was implemented, or how fidelity was confirmed. We could not per- 2655 Records screened
form a meta-analysis owing to a paucity of suitable data, so we
report our results narratively. 2628 Records excluded

Primary Outcomes 27 Full-text articles assessed


None of the included studies provided sufficient data on cardiovas- for eligibility
cular-associated mortality. Four studies15,18,19,21 provided informa-
tion on all-cause mortality, but this information lacked enough data 18 Full-text articles excluded,
with reasons
for meta-analysis. Among inpatient studies, Aliti et al15 (n = 75) re-
ported no deaths in either arm of the trial during the study period;
Velloso et al21 (n = 32) reported 1 death in each arm of the study 9 Studies included in qualitative
synthesis
(owing to sudden death and septic shock after pulmonary infec-
tion), but stated that the deaths were not associated with the in-
0 Studies included in quantitative
tervention. Among outpatient studies, Philipson et al19 (n = 30) synthesis (meta-analysis)
reported 1 death in each arm of the study, and Colín-Ramírez et al18
(n = 38) reported 1 death in the control arm. Neither author group Databases searched included the Cochrane Central Register of Controlled Trials,
specified the cause of death. None of the included studies pro- MEDLINE, Embase, CINAHL, and ClinicalTrials.gov to February 29, 2016;
an updated search of Cochrane Central Register of Controlled Trials and
vided any outcome data on stroke or myocardial infarction or re-
MEDLINE to March 31, 2018; and current trials identified in ClinicalTrials.gov.
ferred directly to the presence of hyponatremia or hypernatremia.

Secondary Outcomes tional class). The investigators reported no significant between-


Hospitalization group differences in the changes in the Clinical Congestion Score from
Among the inpatient studies, Aliti et al15 (n = 75) included patients baseline to 3-day reassessment (primary end point) between the
with acute illness and reported no significant between-group dif- intervention group (mean [SD], –4.03 [3.30] points) and the con-
ferences in the number of hospital readmissions within 30 days of trol group (mean [SD], –3.44 [3.35] points; P = .47). The authors
the end of the study, which ran for 7 days (intervention group, 11 pa- reported that both groups had similar improvements in clinically
tients [29%]; control group, 7 patients [19%]; P = .41). Among the overt congestion. They followed up 37 patients from the interven-
outpatient studies, Hummel et al22 (n = 66) reported that 30-day tion group and 34 from the control group for 30 days after
readmissions were lower in the group with restriction of sodium discharge. Patients randomized to the intervention group had sig-
intake but provided no specific data. nificantly more congestion than the controls (mean [SD], 7.9 [3.8]
vs 6.0 [3.1] points; P = .02). Velloso et al21 (n = 32) reported no sig-
Length of Hospital Stay nificant difference in the time needed for compensation of heart
Aliti et al15 (n = 75) reported that the median length of stay among failure symptoms between the intervention group (mean [SD],
inpatients was 6 days (range, 4-12), with no significant difference be- 7.5 [1.9] days) and the control group (6.6 [1.6] days; P = .18).
tween the intervention group (median, 7 days; 95% CI, 3.8-13.0 days)
and control group (median, 6 days; 95% CI, 4.0-12.5 days; P = .89). Outpatient Studies | Four studies16-18,20 reported changes in NYHA
They reported no statistically significant differences in length of stay functional class before and after reduced salt intake. Two studies18,22
among the patients who remained hospitalized after day 7, regard- reported changes in clinical condition using the Kansas City Cardio-
less of allocation (P = .90). myopathy Questionnaire (KCCQ) Clinical Summary Score. Alvelos et
al16 (n = 24) reported that NYHA functional class was not improved
Changes in the Clinical Signs and Symptoms of Heart Failure by restriction of salt intake during the 15-day study, but they did not
We identified 7 studies15-19,21,22 that reported clinical changes in signs provide data to justify this statement. Colín-Ramírez et al17 (n = 65)
and symptoms associated with heart failure using a predefined reported that the main signs and symptoms of heart failure were less
method or validated classification system (eg, NYHA functional class). common in the intervention group (before and after the study), with
The main results from these 7 studies are summarized in Table 2. significant differences for edema (37% vs 7.4%; P = .008) and fa-
tigue (59% vs 26%; P = .012). They reported no significant changes
In-Patient Studies | Aliti et al15 (n = 75) reported changes in the Clini- in the control group. The authors also reported that patients in the
cal Congestion Score, an instrument composed of 7 items to assess intervention group had significant improvements in NYHA func-
signs and symptoms of congestion (including rales, a third heart tional class, with fewer patients in NYHA classes II and III and more in
sound, jugular venous distention, peripheral edema, hepatojugular class I by the end of the study (P = .03). This difference was not found
reflux, orthopnea, paroxysmal nocturnal dyspnea, and NYHA func- among participants in the control group.

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine Published online November 5, 2018 E3

© 2018 American Medical Association. All rights reserved.

Downloaded From: on 11/05/2018


E4
Table 1. Summary of Included Studies
No. of Mean Age of
Source Country Participants Population Inclusion Criteria Participants, y Intervention Comparator Duration of Study
Inpatient Studies
Aliti et al,15 Brazil 75 Adult patients with a diagnosis of acute 60 2 g/d salt and 800 mL/d water 7.5-12.5 g/d salt and estimated Outcomes assessed until hospital

Downloaded From: on 11/05/2018


2013 decompensated HF defined as LVEF ≤45%; ≥2500 mL/d water day 7 or until discharge in patients
Boston criteria score of ≥8 points; and LOS with LOS <7 d; some assessments
≤36 h after hospital admission at day 30
Velloso et al,21 Brazil 32 Adult patients admitted to hospital with 54 2 g/d salt ≤10 g/d salt Unclear, although appears time
1991 acute illness due to underlying chronic HF taken for patients to compensate
Outpatient Studies
Clinical Review & Education Review

Alvelos et al,16 Portugal 24 Adult patients with mild to moderate 70 100 mmol/d sodium Maintenance of a diet with usual 15 d
2004 chronic HF (euvolemic) defined as LVEF salt intake
≤40% with no exacerbations or
therapeutic changes in the previous 2 mo
Colín-Ramírez Mexico 65 Adult patients with a confirmed diagnosis 62 Aim of 2.0-2.4 g/d sodium (5-6 g/d salt) Traditional management of HF, 6 mo
et al,17 2004 of HF defined as decreased systolic and/or including common dietary
diastolic function as determined by advisories regarding decreased
ECG criteria sodium and fluid intakes but no
specific prescription
Colín- Ramírez Canada 38 Adult patients with a confirmed diagnosis 66 Aiming for 65 mmol/d or 1.5 g/d sodium Moderate intake of sodium aiming 6 mo
et al,18 2015 of HF (reduced and preserved systolic (3.75 g/d salt) for 100 mmol/d or 2.3 g/d
function), NYHA classes II and III, and (5.75 g/d salt)
receiving optimally tolerated medical
therapy according to CCS guidelines

JAMA Internal Medicine Published online November 5, 2018 (Reprinted)


Hummel et al,22 United 66 Adults aged ≥55 y with history of systemic 72 Daily sodium intake of 1.5 g/2100 kcal; Usual care including pamphlet, 12 wk
2017 States hypertension and acutely decompensated compliant meals for 4 wk after hospital “How to Eat a Low Sodium Diet”
HF as primary diagnosis for admission or discharge in addition to pamphlet “How and telephone call from study
secondary diagnosis after hospitalization to Eat a Low Sodium Diet” and telephone staff at 2 and 3 wk
for another reason followed by discharge call from study staff at 2 and 3 wk
into the community
Philipson Sweden 30 Adult patients with a history of CHF in 74 Aiming for a maximum of 2-3 g/d sodium General diet information in 12 wk
et al,19 2010 NYHA classes II and IVa (5.0-7.5 g/d salt) and to restrict fluids to accordance with ESC guidelines
1.5 L/d for heart failure
Philipson Sweden 97 Adult patients with a history of CHF, in 75 Advice to reduce sodium intake to Dietician- or nurse-led standard 12 wk; Patients were also
et al,20 2013 NYHA classes II and IVb 2-3 g/d (5 g/d salt) and to limit fluid advice (eg, be aware not to drink contacted by telephone after
intake to a maximum of 1.5 L/d too much and use salt with 10-12 mo by a dietician
caution)
Welsh et al,23 United 52 Adult patients who had a confirmed 63 6-wk Education intervention with Usual care and visited at 3 data 6 mo
2013 States diagnosis of HF due to left ventricular instruction and advice on restriction of collections
systolic dysfunction or with preserved sodium diet adherence from the

© 2018 American Medical Association. All rights reserved.


systolic function and NYHA classes II intervention nurse during home visits
and IVc and telephone calls (no set salt goal set)
Abbreviations: CCS, Canadian Cardiovascular Society; CHF, congestive heart failure (HF); ECG, echocardiography; peripheral edema); received maximum tolerated doses of angiotensin-converting enzyme inhibitors or
ESC, European Society of Cardiology; LOS, length of stay; LVEF left ventricular ejection fraction; NYHA, New York angiotensin receptor blockers and β-blockers with no change in medication for at least 2 weeks before
Heart Association; enrollment; and received a dose of 80 mg or more of furosemide or equipotent other diuretics for patients in
a NYHA class II or at least 40 mg of furosemide for patients in NYHA classes III to IV.
Participants must have been in a stable condition with no change in medication for at least 2 weeks before
c
enrollment; have documented LVEF of less than 40% or greater than 40% with a history of hospitalization for Participants were recruited from a cardiology clinic, community hospital, and a university hospital; have had no
HF; had signs of fluid retention; and received doses of furosemide or equipotent diuretics of greater than 80 mg cognitive impairments limiting the ability to complete an interview or engage in the educational intervention;
for NYHA class II or greater than 40 mg for NYHA classes III to IV. were not living in an extended care facility; were 21 years or older; had a hospitalization or emergency
b department treatment of HF 1 year before study entry; lived 90 miles or closer to the hospital or clinic; and
Participants must have been in a stable condition with documented LVEF of no more than 40% or greater than
40% with a history of hospitalization for HF; a history of signs of fluid retention (on chest radiography or spoke English.
Reduced Salt Intake for Heart Failure

jamainternalmedicine.com
Reduced Salt Intake for Heart Failure Review Clinical Review & Education

Figure 2. Risk of Bias Summary

Colín-Ramírez et al,17 2004


Colín-Ramírez et al,18 2015

Philipson et al,19 2010


Philipson et al,20 2013
Hummel et al,22 2017
Alvelos et al,16 2004

Velloso et al,21 1991


Welsh et al,23 2013
Aliti et al,15 2013
Random sequence generation (selection bias) + ? ? + ? + + – +
Allocation concealment (selection bias) ? ? ? + ? ? ? – ?
Blinding of participants and personnel (performance bias) ? – – – – ? ? ? –
Blinding of outcome assessment (detection bias) + ? + + + + + ? ?
Incomplete outcome data (attrition bias) + + ? ? ? ? + ? ? Assessment used the Cochrane risk of
Selective reporting (reporting bias) + ? + + ? + + ? +
bias tool whereby minus indicates a
high risk of bias; plus, low risk of bias;
Other bias + ? ? ? ? ? + ? ?
and question mark, unclear risk
of bias.

75 points, with a median change of 9 points (range, 2-5 points;


Table 2. Summary of Changes in the Clinical Signs and Symptoms
of Heart Failure P = .006) and tended to increase in the group with moderate
sodium intake from 66 to 73 points, with a median change of 6
No. of Difference,
Source Country Participants Intervention vs Control (range, −1 to 15 points; P = .07). However, the authors found no
Inpatient Studies significant difference between the groups (P = .41).
Aliti et al,15 2013a Brazil 75 No difference Philipson et al20 (n = 97) used a composite end point, based on
Velloso et al,21 Brazil 32 No difference whether individual patients had shown deterioration or improve-
1991b
ment or were unchanged. Patients were classified into any 1 of these
Outpatient Studies
3 outcomes based on several changes, including change in NYHA func-
Alvelos et al,16 Portugal 24 No difference
2004c tional class, hospitalization for heart failure, weight change, and leg
Colín-Ramírez Mexico 65 Favors intervention edema. The primary composite end point improved in 51% of the
et al,17 2004c
patients in the intervention group and 16% in the control group
Colín-Ramírez Canada 38 No difference
et al,18 2015c (P = .001). The authors stated that this improvement was mostly
Philipson et al,19 Sweden 97 Favors intervention owing to an improved NYHA class and reduced edema. They re-
2013c
ported that NYHA functional class improved significantly by the end
Colín-Ramírez Canada 38 No difference
et al,18 2015d of the study in those who had been randomized to reduced dietary
Hummel et al,22 United 66 No difference salt intake (P = .01).
2017d States
Hummel et al22 (n = 66) reported that the KCCQ clinical sum-
a
Clinical congestion score used to measure clinical change. mary score improved more in the study arm with restricted sodium
b
Time needed for compensation of heart failure symptoms (predefined) used intake (P = .053). However, the score improved significantly from dis-
to measure clinical change.
c
charge to the 4-week assessment in the intervention arm (P ⱕ .001)
Changes in New York Heart Association functional class used to measure
clinical change.
and the control (usual care) arm (P = .001). No difference occurred
d
Changes in Kansas City Cardiomyopathy Questionnaire clinical summary
in the KCCQ summary score between both study arms (P = .37).
scores used to measure clinical change.
Adherence to Reduced Salt Intake
Adherence to a diet with reduced salt intake, however measured,
Colín-Ramírez et al18 (n = 38) reported no significant differ- was variable. Seven studies included or planned to include urinary
ence in NYHA class between intervention and control groups at the sodium collection.15-17,19-22 Two studies18,23 used food diaries.
end of the study (P = .87), although they stated that the propor- Further details are provided in eResults 2 in the Supplement.
tion of patients whose NYHA class improved tended to be higher in
the group with reduced sodium intake. The authors reported no sig- Changes in Blood Pressure
nificant changes in the frequencies of dyspnea, peripheral edema, Three outpatient studies provided data on changes in blood pres-
and fatigue from baseline to 6 months in either group. The authors sure, although too few data were available for meta-analysis. Alvelos
also conducted a post hoc analysis in which they reclassified et al16 (n = 24) reported a decrease in mean (SD) blood pressure
patients according to the sodium intake achieved at the end of (from 98 [7.7] to 85 [4.6] mm Hg; P < .06) in response to reduced
follow-up (6 months). Following this reanalysis, they reported that dietary salt intake. In contrast, they reported that the control group
the proportion of patients who showed an improvement in NYHA showed no differences in mean (SD) blood pressure (from 114 [9.2]
class tended to be higher in the group with low sodium intake than to 108 [3.9] mm Hg; P = .61) before and after the intervention.
in the group with moderate sodium intake (P = .08). The study also Colín-Ramírez et al18 (n = 38) reported no changes in systolic
reported KCCQ scores. At 6 months, the median KCCQ clinical sum- blood pressure between groups but provided no data. Philipson
mary score increased in the group with low sodium intake by 63 to et al19 (n = 30) also reported no significant difference in mean (SD)

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine Published online November 5, 2018 E5

© 2018 American Medical Association. All rights reserved.

Downloaded From: on 11/05/2018


Clinical Review & Education Review Reduced Salt Intake for Heart Failure

systolic blood pressure between the intervention group (baseline, that reducing sodium intake may reduce the risk for morbidity and
128[19]mmHg;oncompletion, 125[22]mmHg)andthecontrolgroup mortality due to cardiovascular disease.33 However, clinicians and
(baseline, 123 [22.3] mm Hg; on completion, 125 [25.5] mm Hg). policy makers should acknowledge the lack of evidence for this inter-
vention in patients who may be reluctant to restrict salt intake and
the social and economic feasibility of achieving reduced-salt diets in
wider populations.34 Restriction of salt intake for inpatients with acute
Discussion
heart failure has an even weaker evidence base.
To our knowledge, this systematic review represents the most This review also has implications for future research. The over-
up-to-date and comprehensive analysis of current evidence of the all lack of high-quality clinical trials and the heterogeneity of the cur-
effects of reduced salt intake in patients with heart failure. Despite rent evidence base mean that evidence from well-designed, well-
identifying 9 RCTs involving a total of 479 participants, we found performed clinical trials is needed. We note that the full results from
no robust high-quality evidence. current trials, such as the Geriatric Out of Hospital Randomized Meal
We found no clinically relevant data on whether reduced dietary Trial in Heart Failure (GOURMET-HF),35 Study of Dietary Interven-
salt intake affected outcomes such as cardiovascular-associated or tion Under 100 mmol in Heart Failure (SODIUM-HF),36 and Dietary
all-cause mortality, cardiovascular-associated events, hospitaliza- Sodium Intake and Outcomes in Heart Failure (ProhibitSodium)37
tion, or length of hospital stay. Three outpatient studies17,18,20 (eTable 3 in the Supplement), may influence future updates of this
reported a trend toward improvement in the clinical signs and review. Part of the uncertainty around this intervention also arises
symptoms of heart failure with reduced intake of dietary salt. from a lack of understanding of the mechanistic rationale. Several
possible mechanisms have been proposed, including a positive
Consistency With Other Studies effect on sodium-fluid homeostasis,38 lower blood pressure in
Despite the paucity of data, our findings are consistent with the re- individuals with concomitant hypertension,39 and a reduction in left
sults of previous reviews. A Cochrane review on advice to reduce ventricular mass.40 Future research should therefore also include
dietary salt consumption in preventing cardiovascular disease26 hypothesis-testing mechanistic studies, including examination of the
found limited evidence that reduced salt intake may increase deaths effects of restricting salt intake on the renin-angiotensin-
in people with heart failure. However, this conclusion was drawn from aldosterone system and its possible clinical implications.
one of the studies included in the withdrawn systematic review by
DiNicolantonio et al.12 A subsequent Cochrane review update27 Strengths and Limitations
excluded studies from this group because the data were unreli- Our review has several strengths, including the most up-to-date
able. We also excluded studies from this group. synthesis of RCT evidence, to date, on this topic. We conducted the
In a more recent review, Health Quality Ontario28 examined the review according to guidance from the Cochrane Handbook for
effects of restricting salt intake in patients with heart failure. Al- Systematic Reviews of Interventions.14 We worked according to a
though the scope of our review was broader, the findings were gen- prepublished protocol. We conducted a broad search of the litera-
erally consistent. The systematic review of nutritional interventions ture, including several medical databases. We used independent
in heart failure by Abshire et al29 had a broader scope than ours but dual-author selection throughout. The main limitation of this re-
searched a narrower field. They highlighted the current uncertainty view is that although at least 2 reviewers independently checked the
in this field and the contrasting data between RCTs and observa- literature searches, double screening, data entry, and risks of bias,
tional studies. The general limitations of inferring causality from ob- relevant studies may have been missed or interpreted incorrectly.
servational studies in this area has been separately acknowledged.30

Implications
Conclusions
Patients with heart failure are generally advised to reduce salt intake.6
In many cases, this recommendation is based on clinical experience Despite broad advocacy, uncertainty remains about the robustness of
of best current practice.31 For outpatients, we found no significant evi- advice to reduce salt intake in patients with heart failure. In this system-
dence of harms from reduction of dietary salt intake and a trend for aticreview,wefoundlimitedevidenceformodestimprovementinclini-
some clinical improvements such as in NYHA functional class. These cal symptoms from reducing salt intake in outpatients. A paucity of
findings suggest that current best practice should not be changed for robust high-quality evidence was available to support or refute current
this patient group. This suggestion is consistent with other evidence guidance.Well-designed,well-described,andadequatelypoweredstud-
that lower salt intake is associated with minimum health risks32 and ies are needed to reduce uncertainty about the use of this intervention.

ARTICLE INFORMATION Acquisition, analysis, or interpretation of data: Administrative, technical, or material support:
Accepted for Publication: July 23, 2018. Mahtani, Onakpoya, Tierney, Roberts, Nunan. Mahtani, Heneghan, Aronson, Roberts, Nunan.
Drafting of the manuscript: Mahtani, Heneghan, Supervision: Mahtani, Hobbs.
Published Online: November 5, 2018. Onakpoya, Roberts.
doi:10.1001/jamainternmed.2018.4673 Conflict of Interest Disclosures: All authors reported
Critical revision of the manuscript for important receivingfundingfromtheNationalHealthService(NHS)
Author Contributions: Dr Mahtani had full access to all intellectual content: Mahtani, Heneghan, Tierney, National Institute of Health Research School for Primary
the data in the study and takes responsibility for the Aronson, Hobbs, Nunan. CareResearch(NIHR)toconductindependentresearch,
integrityofthedataandtheaccuracyofthedataanalysis. Statistical analysis: Onakpoya, Nunan. including systematic reviews. No other disclosures
Concept and design: Mahtani, Heneghan, Aronson, Obtained funding: Mahtani. were reported.
Hobbs, Nunan.

E6 JAMA Internal Medicine Published online November 5, 2018 (Reprinted) jamainternalmedicine.com

© 2018 American Medical Association. All rights reserved.

Downloaded From: on 11/05/2018


Reduced Salt Intake for Heart Failure Review Clinical Review & Education

Funding/Support:ThisworkwassupportedbytheNHS 11. LucanSC.Attemptingtoreducesodiumintakemight DescriptionandReplication(TIDieR)checklistandguide.


NIHRSchoolforPrimaryCareResearch(project222).Drs do harm and distract from a greater enemy. Am J Public BMJ. 2014;348:g1687. doi:10.1136/bmj.g1687
Mahtani, Heneghan, Onakpoya, and Hobbs are also Health. 2013;103(2):e3. doi:10.2105/AJPH.2012.301105 26. Taylor RS, Ashton KE, Moxham T, Hooper L,
coinvestigatorsintheEvidenceSynthesisWorkingGroup 12. DiNicolantonio JJ, Di Pasquale P, Taylor RS, Ebrahim S. Reduced dietary salt for the prevention
(project390),whichissupportedbytheNIHRSPCR.This Hackam DG. Retraction. Low sodium versus normal of cardiovascular disease. Cochrane Database Syst
study was also supported by Clinical Lecturer Award sodium diets in systolic heart failure: systematic Rev. 2011;(7):CD009217.
CL-2011-13-006 from the NIHR (Dr Mahtani) and the review and meta-analysis. Heart. 2013;99(11):820.
Royal College of General Practitioners Scientific 27. Adler AJ, Taylor F, Martin N, Gottlieb S, Taylor RS,
doi:10.1136/heartjnl-2012-302337
Foundation Board (Dr Mahtani). Dr Hobbs was Ebrahim S. Reduced dietary salt for the prevention of
13. VanHornL,YancyC.Dietpreventionandtherapyfor cardiovascular disease. Cochrane Database Syst Rev.
supported by NIHR Collaboration for Leadership
heart failure? Circ Heart Fail. 2013;6(6):1109-1111. 2014;(12):CD009217.
in Health Research and Care Oxford, NIHR Oxford
doi:10.1161/CIRCHEARTFAILURE.113.000868
BiomedicalResearchCentre,OxfordUniversityHospitals 28. Health Quality Ontario. Sodium restriction in
NHS Foundation Trust, and NIHR Oxford Medtech and 14. Higgins JPT, Green S, eds. Cochrane Handbook heart failure: a rapid review. http://www.hqontario
In-Vitro Diagnostics Co-operative. for Systematic Reviews of Interventions Version 5.1 .ca/Portals/0/Documents/evidence/rapid-reviews
[updated March 2011]. The Cochrane Collaboration, /qbp-chf-clinic-sodium-restriction-20141211-en.pdf.
Role of the Funder/Sponsor: The funding sources had
2011. http://handbook-5-1.cochrane.org. Accessed February 2015. Accessed March 1, 2018.
noroleinthedesignandconductofthestudy;collection,
March 1, 2018.
management, analysis, and interpretation of the data; 29. Abshire M, Xu J, Baptiste D, et al. Nutritional
preparation, review, or approval of the manuscript; and 15. Aliti GB, Rabelo ER, Clausell N, Rohde LE, Biolo A,
interventions in heart failure: a systematic review of
decision to submit the manuscript for publication. Beck-da-Silva L. Aggressive fluid and sodium restriction
the literature. J Card Fail. 2015;21(12):989-999.
Disclaimer: The views expressed in this review in acute decompensated heart failure: a randomized
doi:10.1016/j.cardfail.2015.10.004
represent the views of the authors and not clinical trial. JAMA Intern Med. 2013;173(12):1058-1064.
doi:10.1001/jamainternmed.2013.552 30. Wong MMY, Arcand J, Leung AA, Thout SR,
necessarily those of the host institution, the NHS,
CampbellNRC,WebsterJ.Thescienceofsalt:aregularly
the NIHR, or the Department of Health. 16. AlvelosM,FerreiraA,BettencourtP,etal.Theeffect
updated systematic review of salt and health outcomes
Additional Contributions: Liz Spencer, PhD, University of dietary sodium restriction on neurohumoral activity
(December 2015-March 2016). J Clin Hypertens
of Oxford, and Thanusha Ananthakumar, MBBS, and renal dopaminergic response in patients with heart
(Greenwich). 2017;19(3):322-332. doi:10.1111/jch.12970
University of Oxford, contributed to an earlier stage failure. Eur J Heart Fail. 2004;6(5):593-599. doi:10.1016
in the review process and were not compensated /j.ejheart.2003.11.020 31. Scottish Intercollegiate Guidelines Network.
for their contributions. SIGN 147: management of chronic heart failure.
17. Colín-Ramírez E, Castillo Martínez L, Orea Tejeda A,
http://www.sign.ac.uk/sign-147-management-of
Rebollar González V, Narváez David R, Asensio Lafuente
REFERENCES -chronic-heart-failure.html. Accessed April 8, 2018.
E. Effects of a nutritional intervention on body
1. Al-Mohammad A, Mant J. The diagnosis and composition,clinicalstatus,andqualityoflifeinpatients 32. HenneyJE,TaylorCL,BoonCS;InstituteofMedicine
management of chronic heart failure: review with heart failure. Nutrition. 2004;20(10):890-895. (US)CommitteeonStrategiestoReduceSodiumIntake,
following the publication of the NICE guidelines. Heart. doi:10.1016/j.nut.2004.06.010 eds. Strategies to Reduce Sodium Intake in the United
2011;97(5):411-416. doi:10.1136/hrt.2010.214999 States. Washington, DC: National Academies Press
18. Colín-Ramírez E, McAlister FA, Zheng Y, Sharma S,
(US); 2010.
2. Forouzanfar MH, Moran A, Phillips D, et al. Armstrong PW, Ezekowitz JA. The long-term effects of
Prevalence of heart failure by cause in 21 regions: dietary sodium restriction on clinical outcomes in 33. Newberry SJ, Chung M, Anderson CAM, et al.
global burden of diseases, injuries and risk factors— patients with heart failure: the SODIUM-HF (Study of Sodiumandpotassiumintake:effectsonchronicdisease
2010 study. J Am Coll Cardiol. 2013;61(10)(suppl): Dietary Intervention Under 100 mmol in Heart Failure): outcomes and risks. Comparative Effectiveness Review
E7786. doi:10.1016/S0735-1097(13)60786-X a pilot study. Am Heart J. 2015;169(2):274-281.e1. No. 206. Effective Health Care Program. https:
3. Mosterd A, Hoes AW. Clinical epidemiology of doi:10.1016/j.ahj.2014.11.013 //effectivehealthcare.ahrq.gov/topics/sodium-potassium
heart failure. Heart. 2007;93(9):1137-1146. /final-report-2018. June 2018. Accessed March 1, 2018.
19. Philipson H, Ekman I, Swedberg K, Schaufelberger
doi:10.1136/hrt.2003.025270 M. A pilot study of salt and water restriction in patients 34. Wilson N, Nghiem N, Foster RH. The feasibility of
4. Savarese G, Lund LH. Global public health with chronic heart failure. Scand Cardiovasc J. 2010;44 achieving low-sodium intake in diets that are also
burden of heart failure. Card Fail Rev. 2017;3(1):7-11. (4):209-214. doi:10.3109/14017431003698523 nutritious,low-cost,andhavefamiliarmealcomponents.
doi:10.15420/cfr.2016:25:2 PLoS One. 2013;8(3):e58539. doi:10.1371/journal.pone
20. Philipson H, Ekman I, Forslund HB, Swedberg K,
5. Stewart S, Jenkins A, Buchan S, McGuire A, .0058539
Schaufelberger M. Salt and fluid restriction is effective
Capewell S, McMurray JJ. The current cost of heart 35. ClinicalTrials.gov. Geriatric Out of Hospital
in patients with chronic heart failure. Eur J Heart Fail.
failure to the National Health Service in the UK. Eur Randomized Meal Trial in Heart Failure (GOURMET-HF).
2013;15(11):1304-1310. doi:10.1093/eurjhf/hft097
J Heart Fail. 2002;4(3):361-371. doi:10.1016 NCT02148679. https://clinicaltrials.gov/ct2/show
/S1388-9842(01)00198-2 21. Velloso LG, Alonso RR, Ciscato CM, Barretto AC,
/NCT02148679. Accessed July 11, 2018.
6. GuptaD,GeorgiopoulouVV,KalogeropoulosAP,etal. Bellotti G, Pileggi F. Diet with usual salt allowance in
hospital treatment of congestive heart failure [in 36. ClinicalTrials.gov. Study of Dietary Intervention
Dietary sodium intake in heart failure. Circulation. 2012;
Portuguese]. Arq Bras Cardiol. 1991;57(6):465-468. Under 100 mmol in Heart Failure (SODIUM-HF).
126(4):479-485. doi:10.1161/CIRCULATIONAHA.111
NCT02012179. https://clinicaltrials.gov/ct2/show
.062430 22. HummelSL,KarmallyW,GillespieBW,etal.Geriatric
/NCT02012179. Accessed July 11, 2018.
7. Yancy CW. The uncertainty of sodium restriction in Out-of-Hospital Randomized Meal Trial in Heart Failure
heart failure: we can do better than this. JACC Heart Fail. (GOURMET-HF) pilot randomized trial: primary results. 37. ClinicalTrials.gov. Dietary Sodium Intake and
2016;4(1):39-41. doi:10.1016/j.jchf.2015.11.005 J Card Fail. 2017;23(11):830. doi:10.1016/j.cardfail.2017 Outcomes in Heart Failure (ProhibitSodium).
.10.005 NCT02467296. https://clinicaltrials.gov/ct2/show
8. Butler J, Papadimitriou L, Georgiopoulou V,
/NCT02467296. Accessed July 11, 2018.
Skopicki H, Dunbar S, Kalogeropoulos A. Comparing 23. Welsh D, Lennie TA, Marcinek R, et al.
sodium intake strategies in heart failure: rationale Low-sodium diet self-management intervention in 38. He FJ, MacGregor GA. Reducing population salt
and design of the Prevent Adverse Outcomes in heart failure: pilot study results. Eur J Cardiovasc Nurs. intake worldwide: from evidence to implementation.
Heart Failure by Limiting Sodium (PROHIBIT) Study. 2013;12(1):87-95. doi:10.1177/1474515111435604 Prog Cardiovasc Dis. 2010;52(5):363-382. doi:10.1016
Circ Heart Fail. 2015;8(3):636-645. doi:10.1161 /j.pcad.2009.12.006
24. Wessler JD, Maurer MS, Hummel SS. Evaluating the
/CIRCHEARTFAILURE.114.001700 39. He FJ, Li J, Macgregor GA. Effect of longer term
safety and efficacy of sodium-restricted/Dietary
9. DiNicolantonio JJ, Di Pasquale P, Taylor RS, Approaches to Stop Hypertension diet after acute modest salt reduction on blood pressure: Cochrane
Hackam DG. Low sodium versus normal sodium decompensatedheartfailurehospitalization:designand systematic review and meta-analysis of randomised
diets in systolic heart failure: systematic review and rationale for the Geriatric Out of Hospital Randomized trials. BMJ. 2013;346:f1325. doi:10.1136/bmj.f1325
meta-analysis. [published online March 12, 2013]. Meal Trial in Heart Failure (GOURMET-HF). Am Heart J. 40. Ferrara LA, de Simone G, Pasanisi F, Mancini M,
Heart. 2013. doi:10.1136/heartjnl-2012-302337 2015;169(3):342-348.e4. doi:10.1016/j.ahj.2014.11.021 Mancini M. Left ventricular mass reduction during
10. Jun M, Neal B. Low dietary sodium in heart salt depletion in arterial hypertension. Hypertension.
25. Hoffmann TC, Glasziou PP, Boutron I, et al. Better
failure: a need for scientific rigour. Heart. 2014;100 1984;6(5):755-759. doi:10.1161/01.HYP.6.5.755
reporting of interventions: Template for Intervention
(21):e2. doi:10.1136/heartjnl-2012-303266

jamainternalmedicine.com (Reprinted) JAMA Internal Medicine Published online November 5, 2018 E7

© 2018 American Medical Association. All rights reserved.

Downloaded From: on 11/05/2018