Salt and Fluid Restriction Is Effective in Patients With Chronic Heart Failure

Henriette Philipson; Inger Ekman; Heléne B. Forslund; Karl Swedberg; Maria Schaufelberger


Eur J Heart Fail. 2013;15(11):1304-1310. 

In This Article


In this study, salt and fluid restrictions were associated with a significant improvement in a composite endpoint, comprising NYHA class, hospitalization for heart failure, weight, leg oedema, QoL, thirst, and lowered dose of diuretics, in patients with CHF. We believe the effects of the intervention were obtained because concerted efforts were made to meet the nutritional needs of the individual and to promote a good nutritional status with an individualized dietary plan.

Salt and fluid restriction has long been recommended for patients with CHF in national and international guidelines,[1,2] although the scientific evidence has to date been weak. In the 2012 ESC guidelines, the salt and fluid restrictions have been modified to 'avoid excessive fluid intake', 'weight-based fluid restriction may cause less thirst', and salt restriction is listed in 'Gaps in evidence'.[3]

However, the present study gives support to this recommendation. The earlier lack of evidence is understandable as salt and fluid intake is difficult to assess reliably, especially in an outpatient setting. In addition, it can be hard for this patient group, i.e. elderly, ailing patients often on diuretics, to follow diet instructions. Hence, we conducted a pilot study[12] in advance of this study to verify the feasibility of the study procedure. The present study adds to our knowledge about the efficacy of salt and fluid restriction.

Fluid Restriction

Fluid restriction of 1.5 L daily has long been recommended for patients with CHF in guidelines.[1–3] To our knowledge, there are only a few randomized trials of fluid restriction in patients with HF. For example, in a small, in-hospital study comparing a fluid-restricted diet (1 L/day) with free fluid intake, no differences were found in days of hospitalization, time to discontinuation of i.v. diuretics, or stabilization of urea, creatinine, or potassium. The authors therefore conclude that fluid restriction has no clinical benefit for this group of patients.[21] In an outpatient setting, Holst and co-workers compared two different fluid restriction regimens and found that liberal fluid intake (30–35 mL/kg body weight/day) was not significantly better than a restricted fluid intake (1.5 L/day) with respect to body weight, signs and symptoms, diuretic use, QoL, or physical capacity.[9] However, patients with the more liberal fluid regimen complained less about thirst.[22] In our study, only a few patients complained of thirst, possibly owing to the fact that we made a point of openly exchanging ideas with the patients about how best to spread out fluid intake over the course of the day.

Salt and Fluid Restriction

In a study using the same salt and fluid restrictions as ours but with a more complicated diet, Colin Ramirez and co-workers[4] showed that decreased fluid intake and urinary sodium excretion results in less frequent oedema and fatigue, and improved NYHA class and QoL.[4] These results are in accordance with our findings, although their patients were 'healthier' than ours in the sense that nearly 60% of the patients in that study were NYHA class I already at baseline. In addition, the patients were relatively small, 64 (±16) kg and 1.52 (± 0.11) m compared with the averages for persons in Western countries. Furthermore, the lack of control of real fluid intake and urinary excretion makes it difficult to evaluate adherence to the dietary restrictions.

Salt Restriction

Guideline recommendations for salt intake in patients with CHF vary from the vague 'excessive intake of salt should be avoided'[1] to the specific 'patients should be educated to limit their salt intake to a maximum 2–3 g sodium daily'.[23] In line with the latter, a salt restriction of 2–2.4 g sodium daily has previously been shown to decrease urinary sodium excretion and improve NYHA class and QoL.[4] This finding was confirmed in our study. On the other hand, another study showed that patients with furosemide doses of 250–500 mg t.i.d., a fluid intake of 1 L/day, and a sodium restriction of 80 mmol/day had a worse outcome than those on 120 mmol/day.[5] That study differs from ours in several ways, the most exceptional difference being that their patients were on high doses of diuretics and low fluid intake. Hence, we agree with the authors that their results must be interpreted with caution.[5] In contrast, a recent prospective study has shown that patients with CHF and a high sodium diet have a greater risk of acute decompensated heart falure, all-cause hospitalization, and mortality.[24]

One of the challenges in research studies on salt restriction is to assess adherence to a low salt diet effectively, which was done by means of urine collection in both Colin Ramirez et al.[4] and the Heart Failure Society of America study.[22] To ensure that the urine collection is complete, it is essential to use a marker in the urine. In our case, we selected PABA as the marker for urine collection completeness. The use of urine PABA concentrations also enables sodium intake to be estimated when urine collections are incomplete. To our knowledge, this is the best method available today.[13,14]

Dietary Recommendations

Prochaska et al.[25] maintain that lifestyle changes, and especially diet changes, are difficult to realize in elderly patients.[26,27] In this study, the dietary recall procedure and the modified semi-quantitative FFQ[18] were used to optimize nutrition counselling and to gain a comprehensive understanding of the patients' personal eating habits. The counselling was tailored to each person and discussed in such a way that the patients were able to adhere to the new diet without negative impacts on QoL. To support and maintain the dietary changes, unannounced telephone interviews were conducted using the 24 h dietary recall procedure, which also served as input for additional guidance and dietary instruction.[20]

Since sodium is unevenly distributed in the everyday diet and is influenced by our priorities, which, in turn, are influenced by cultural and culinary traditions, the generalizability of our results has to be made with some caution. However, we believe that the individualization of change of diet, which was done in this study, has to be one cornerstone of the therapy, and this can be done in all cultural settings, making the results generalizable.


First, most of the patients in both groups had already received dietary instructions in accordance with CHF guidelines prior to inclusion.[1,2] Although it is common practice today to give CHF patients such advice, we were careful to ensure that the control group received standard dietary instructions. It is conceivable that the observed effects of salt and fluid restriction would have been greater if the patients had not received prior dietary advice. Secondly, it is impossible to perform a double blind intervention diet study out of hospital. To limit this bias, the physicians who examined the patients at baseline and follow-up were blinded to the randomization. Thirdly, all patients were on diuretics, which can interfere with the salt and water balance; however, doses were changed only in a few patients and we therefore do not consider that this affected our results. Fourthly, the study had a limited follow-up of 12 weeks. A longer follow-up would most probably have led to more events and more power. In addition, at the 1-year telephone follow-up, more than half of the patients had also maintained the dietary pattern after the end of the study. The limited power is probably the reason for the discrepancy between the positive findings for NYHA class and oedema, while the components of QoL, e.g. thirst, were neutral. Seven patients (7%) dropped out of the study, of which two patients died. Hence, only five patients were not willing or able to finish the study. The few drop-outs, which is a strength not a limitation, shows that the regimen chosen in the intervention was possible to follow. Fifthly, even for healthy subjects, it is difficult to remember to sample all urine, and our patients were old and sick, thus it was even more difficult for them to remember to collect their urine and also to get the complete amount of urine into the container. However, even if there are weaknesses with this method, it is to our knowledge the best one available today in outpatients, especially as we gave PABA to the patients during one of the 24 h urine sampling periods, Sixthly, renal function is often decreased in the elderly and may interfere with PABA excretion. Nevertheless, it is essential to verify the completeness of the urine collection, and we believe that assessing PABA recovery is the best and most appropriate method for this purpose.

Finally, we did not reach the calculated number of patients since the study became too extended in time. However, even if the study was underpowered, it reached statistical significance in the pre-defined composite endpoint. Therefore, we believe the results are true and add information to earlier knowledge in the field. The limited power is probably the reason for the discrepancy between the positive findings for NYHA and oedema, while the components of QoL, e.g. thirst, were neutral.