Ikenoue T, Koike K, Fukuma S, Ogata S, Iseki K, Fukuhara S: Salt Intake and All-Cause Mortality in Hemodialysis Patients. Am J Nephrol 2018;48:87–95
The relationships between sodium chlorid (NaCl) intake and clinical outcomes in hemodialysis patients (such as mortality rates) are difficult to study and controversial. Urinary NaCl excretion rate, a valid method to estimate daily consumption of NaCl in subjects with normal renal function, cannot be used in subjects with advanced chronic kidney disease (CKD) receiving hemodialysis (HD; especially in anuric subjects).
Ikenoue and colleagues conducted a retrospective, registry-based (hypothesis generating) study in 88,116 Japanese prevalent HD patients in order to determine the association of estimated (not measured) NaCl intake with all cause (cancer excluded) and cardiovascular disease (CVD) mortality. NaCl intake was estimated from intra-dialytic weight loss as a surrogate for inter-dialytic weight gain. The follow-up for determining the selected outcomes was only one year. About 34 % of the patients included in the study had missing data. The data points were imputed in the study design in these patients.
Surprisingly, at least to me, estimated NaCl intake in the low range (< 6 gm NaCl/d) was associated with the highest all-cause mortality (ACM), even after adjustment for multiple confounding variables, except for smoking. The lowest risk for ACM was observed at a NaCl intake of 9 gm/d. There was a “hint” of a “U” shaped curve for ACM but not CVD mortality in the comparisons with estimated NaCl intake.
While rigorous in design and analysis, this study cannot implicate low NaCl intake as a causal factor in enhancement of ACM or CVD mortality rates. In addition, the follow-up times for calculation of event rates were quite short (only one year). Furthermore, as this was a registry-based study, it was not possible to validate the accuracy of the NaCl intake estimates by use of intra-dialytic weight loss. The exclusively Japanese character of this study limits its application to other populations. The high “missing” data points are also of concern.
While current guidelines concerning ideal NaCl intake for HD patients suggest values of < 6 gm/d no interventional controlled studies have demonstrated that these values are optimal. This study, despite its flaws, provides some weak but tantalizing evidence that low NaCl intake may be harmful. But, I would suspend judgement on this conclusion until interventional trials are executed using direct measurement of NaCl intake (by dietary history recall) comparing longer term outcomes with high and low NaCl intake. This would not be an easy study to conduct, due to effect of varying NaCl intake on inter-dialytic weight gain and the need for aggressive ultrafiltration during dialysis session. But at least it might shed some light on the validity of using inter-dialytic weight loss as an estimate of NaCl intake, which is at the core of this hypothesis generating study.