Sodium intake is associated with increased disease activity in multiple sclerosis

Abstract

Background

Recently, salt has been shown to modulate the differentiation of human and mouse Th17 cells and mice that were fed a high-sodium diet were described to develop more aggressive courses of experimental autoimmune encephalomyelitis. However, the role of sodium intake in multiple sclerosis (MS) has not been addressed. We aimed to investigate the relationship between salt consumption and clinical and radiological disease activity in MS.

Methods

We conducted an observational study in which sodium intake was estimated from sodium excretion in urine samples from a cohort of 70 relapsing-remitting patients with MS who were followed for 2 years. The effect of sodium intake in MS disease activity was estimated using regression analysis. We then replicated our findings in a separate group of 52 patients with MS.

Results

We found a positive correlation between exacerbation rates and sodium intake in a multivariate model adjusted for age, gender, disease duration, smoking status, vitamin D levels, body mass index and treatment. We found an exacerbation rate that was 2.75-fold (95% CI 1.3 to 5.8) or 3.95-fold (95% CI 1.4 to 11.2) higher in patients with medium or high sodium intakes compared with the low-intake group. Additionally, individuals with high-sodium intake had a 3.4-fold greater chance of developing a new lesion on the MRI and on average had eight more T2 lesions on MRI. A similar relationship was found in the independent replication group.

Conclusions

Our results suggest that a higher sodium intake is associated with increased clinical and radiological disease activity in patients with MS.

INTRODUCTION

Multiple sclerosis (MS) is an inflammatory demyelinating disease caused by an autoimmune response against the central nervous system that is presumed to result from a complex interplay between genes and the environment.¹

The autoimmune response directed against the central nervous system involves several immune cell types. Of particular interest is the Th17 T-cell subset, which has gained particular attention in recent years.²

Studies showing a shifting incidence of MS in migrant populations and its geographical distribution indicate that environmental factors play a major role in disease pathogenesis. Consequently, it has been demonstrated that several environmental factors, such as vitamin D,³ Epstein-Barr virus infection,⁴ systemic infections,⁵ parasitical infections⁶ and immunisations,⁷ might play a role in the onset, as well as the clinical, radiological and immunological courses of MS.

Recently, salt (NaCl, sodium chloride) has been shown to modulate the differentiation of human and mouse Th17 cells.^{8 9} Moreover, mice that were fed a high-sodium diet were described to develop more aggressive courses of experimental autoimmune encephalomyelitis (EAE), the animal model of MS.^{8 9}

However, the role of sodium intake in MS has not been addressed. In this study, we investigated the effect of sodium intake on the clinical and radiological activity of MS.

MATERIALS AND METHODS

Study population and design

The study was approved by the Institutional Ethics Committee, and all participants signed an informed consent form. The study included two participants of patients: group 1 and group 2 (figure 1). The first group consisted of a traditional cohort to test our main hypothesis, and the second group was made up of a cross-sectional sample recruited for replication purposes. Group 1 consisted of 70 patients with relapsing–remitting MS according to the McDonald criteria¹⁰ who were seen at the MS clinic at the Raúl Carrea Institute for Neurological Research and were consecutively invited to participate in a longitudinal cohort study in which an MRI was required at baseline and clinical, radiological and sodium intake data were collected throughout the follow-up period (from September 2010 until November 2012). Group 2 consisted of a cross-sectional sample of 52 relapsing–remitting patients with MS according to the McDonald criteria who were recruited from June to July 2013. Pregnant patients and patients with hyponatraemia were excluded from enrolment in both groups.

Figure 1.

Study design. Two subsets of patients were included in this study, group 1 and group 2. The first group consisted of a cohort of 70 patients recruited in September 2010; this cohort was followed for two years. All patients had a baseline MRI and an MRI at least yearly. At the middle of the follow-up, coinciding with the yearly MRI, a morning first-urine sample was taken and sodium intake was calculated. Neurological assessments were performed every 3–6 months until the end of the follow-up. The main outcome was the number of relapses at the end of follow-up with radiological activity as secondary outcomes. The second group was recruited as a replication group, casual urine sample was taken, and the T2 lesion load was the comparator outcome.

Blood and urine samples were taken 12 months after enrolment for patients in group 1. Sodium and creatinine levels in urine and serum sodium and vitamin D levels were measured via routine analysis in our hospital’s clinical laboratory. For group 2, casual urine samples were collected and analysed by following the same procedures as those used for the initial cohort.

The relationship between salt consumption and MS disease activity was tested by applying cross-sectional and longitudinal analyses. In the former, correlations between T2 lesion load and sodium intake were investigated. This analysis was performed first in group 1 (at month 12 when a urine sample was taken and an MRI was performed) and then repeated for validation in group 2. The longitudinal cohort (group 1) follow-up lasted 2 years (sample taken at month 12), and the association between salt consumption and increased clinical and radiological activity during follow-up was tested.

Sodium intake assessment

We estimated sodium excretion in urine as a proxy of sodium intake because 80–90% of ingested sodium is eliminated through urine.¹¹ Twenty-four hour urine collection is the most reliable method for measuring sodium excretion and therefore intake but collecting 24 h urine is cumbersome and can lead to inaccurate measurement due to patients skipping some urine collection throughout the day.¹² Thus, alternative methods have been developed to estimate sodium excretion; Tanaka’s equation uses excreted sodium and creatinine in spot urine to estimate the total sodium excretion,¹³ has been widely assessed in the general population, and has been recommended as a reliable and convenient method for calculating salt intake via hypertension guidelines.¹⁴

Tanaka’s equation is as follows:

$$21.98\times {(\text{urinary}\text{sodium}/\text{urinary}\text{creatinine}\times (-2.04\times \text{age}+14.89\times \text{weight}+16.14\times \text{height}-2244.45))}^{0.392}.$$

Patients were asked to provide urine from their first morning urination on a weekday. All patients were required to provide urine samples after 3, 6 and 9 months to assess potential changes in salt consumption with time and to test method reliability. Patients in group 2 were asked to provide urine on the same day of recruitment regardless of time or number of previous voidings that day. Urine samples were separated at least for 90 days since last steroid administration.

Sodium intake was then divided into three levels according to the recommendations of the WHO as the target to avoid high-blood pressure-related disorders: patients with less than 2 g/day, patients with an average intake according to national standards (between 2 and 4.8 g/day) and patients whose intake was above average (more than 4.8 g).¹⁵

Clinical data

Clinical data were retrieved from our patient with MS database. The Expanded Disability Status Scale (EDSS) score was assessed at baseline and at the end of follow-up. The number of relapses during follow-up was counted during this period. Additionally, disease duration was calculated from the date of diagnosis until the end of follow-up. Other features extracted were gender, age, treatment history, body weight, height, vitamin D level (measured at the same timepoints as sodium intake) and smoking status (measured at baseline).

Exacerbation was defined as development of a new symptom or worsening of pre-existing symptoms confirmed via neurological examination, lasting at least 48 h and preceded by stability or improvement lasting at least 30 days.

MRI assessment

Brain and spinal cord MRI was performed at baseline and at 3-month or 6-month follow-up visits on a 1.5-T MRI unit (General Electric, Milwaukee, Wisconsin, USA). Axial 5 mm-thick slices with no gap and a 192×256 matrix with sub-callosum alignment were obtained with T2-weighted, proton density, fast spin-echo, fluid-attenuated inversion recovery and T1-weighted sequences prior to and after the administration of gadolinium (0.1 mmol/kg).

In addition to the neuroradiologist report, a blinded assessment of MRI was performed by JC and MFF. T2 lesions and gadolinium-enhancing lesions were noted. Additionally, the combined unique activity (CUA) was calculated by adding new or enlarging T2 lesions to the gadolinium-enhancing lesions as previously reported.¹⁶ Conflicts in MRI assessment were solved by discussion between JC and MFF.

Statistical analysis

The primary outcome of the study was the number of relapses occurring from the time of recruitment to the time of the last follow-up. Secondary outcomes were the CUA and the T2 lesion load on MRI. The primary predictor for the analysis was the estimated sodium intake.

Demographical, clinical and radiological baseline characteristics between the low, medium and high salt-consuming groups were compared using the Kruskal-Wallis test and χ² when appropriate.

A Poisson regression model was used to assess the impact of sodium intake on the number of clinical relapses, generating an incidence rate ratio and corresponding 95% CIs. Fitness of the Poisson regression model over the zero-inflated Poisson model was tested with the Vuong test. Unadjusted and adjusted values were calculated for all analyses including potential confounders such as age (salt intake may vary with age), gender (there are reported differences in sodium intake with gender), disease duration (increase lesion load occurring over time may confound the association), treatment (there may be an unknown association between treatment and dietary preferences, untreated and patients with immunomodulatory drugs were grouped and used as baseline in comparison to patients treated with immunosuppressive drugs), vitamin D levels (there may be a correlation between sodium intake and vitamin D levels), body mass index (BMI, obesity has been linked to sodium intake) and smoking status (smoker, non-smoker, we included this because it may be related to health consciousness). The effects of salt intake on EDSS progression (dichotomised in individuals who increased their score and individuals who maintained or improved their score) were analysed via logistic regression. The association between sodium consumption and T2 lesion load was evaluated using the same method for group 1 and 2 with an adjusted linear regression model with heteroskedasticity-robust SDs. All statistical analyses were performed using Stata V.12 software.

RESULTS

Clinical and demographic characteristics of the study participants

The clinical and demographic characteristics of the study participants are presented in table 1 and online supplementary table. A total of 70 patients were included in the initial cohort, and 52 were included in group 2; thus, a total of 122 patients were included. For the initial cohort, the mean follow-up time was 24 months (range 22–26 months). None of the patients suffered from hypertension or had been prescribed diuretics. None of the patients were on vitamin supplementation, and four patients started multivitamin supplements during follow-up. None of the patients were on a weight-loss diet. The 70 initial cohort participants experienced a total of 44 relapses with an annualised relapse rate of 0.31. The replication group was similar to the initial cohort in terms of age (p=0.41), disease duration (p=0.42) and EDSS (p=0.21); however, the replication group had more males (p=0.007; see online supplementary table).

Table 1.

Baseline and clinical characteristics of the study participants

	All participants (n=70)	Low-sodium intake (n=21)	Medium sodium intake (n=37)	High-sodium intake (n=12)	p Value
Age (years, means±SD)	37.5±8.9	42.27±10.1	35.66±8.3	40.6±8.9	0.73
Female:male (n)	54:16	20:1	28:9	6:6	0·05
Disease duration (years, median, range)	5 (1–16)	4 (1–16)	3 (1–16)	4 (1–14)	0.18
EDSS (median, range)	1 (0–3·5)	1 (0–3)	0 (0–3.5)	0 (0–2.5)	0.83
Treatment (n)
None	1	0	0	1
Interferon	30	11	15	4
Glatiramer acetate	22	8	9	5
Natalizumab	4	0	4	0
Fingolimod	13	2	9	2

EDSS, Expanded Disability Status Scale.

Estimated sodium intake

The mean estimated daily sodium intake for the entire population was 4.12±1.6 g/day, which is in the range estimated for Argentina (4–4.8 g). As previously reported,¹¹ males had significantly higher levels of sodium intake compared with females (5.3±1.8 vs 3.78±1.4 (p<0.001)). There were no significant differences in sodium intake between treatment groups (p=0.45).

Similar results were found for the replication group, which had an average daily sodium intake of 4.45±1.5 (p=0.24), and no differences between treatment groups (p=0.34) were observed. Unlike the initial cohort, males from the replication group did not significantly differ from females in terms of salt intake (p=0.87).

Sodium intake and clinical outcomes

The relationship between exacerbation rate and sodium intake levels is shown in table 2. To assist in analysing the association between sodium intake and clinical activity, sodium intake was categorised into three groups: a baseline group comprising patients with the WHO recommended intake (under 2 g/day), patients within the national average (2–4.8 g) and patients with an above average intake (4.8 g or more).

Table 2.

Association between sodium intake and exacerbation rate in a regression analysis

	IRR	95% CI	p Value
IRR of exacerbation (univariate model)
Sodium intake (g/day)
<2	1 (baseline)	–	–
2–4.8	2.56	1.3 to 4.9	0.005
>4.8	3.37	1.5 to 9.55	0.001
IRR of exacerbation (adjusted model)
Sodium intake (g/day)
<2	1 (baseline)	–	–
2–4.8	2.75	1.3 to 5.8	0.008
>4.8	3.95	1.4 to 11.2	0.01
Age (1-year increment)	0.992	0.96 to 1.02	0.59
Gender (male)	1.09	0.49 to 2.42	0.82
Disease duration (1-year increment)	0.99	0.98 to 1.01	0.08
Vitamin D (1 ng increase)	1	0.96 to 1.04	0.85
Smoking (smoker)	1.13	0.56 to 2.28	0.73
BMI (1 unit increase)	0.97	0.87 to 1.07	0.58
Treatment (immunosuppressant vs immunomodulators/untreated)	1.46	0.79 to 2.73	0.22

BMI, body mass index; IRR, incidence rate ratio.

We found a positive correlation between exacerbation rate and sodium intake in a multivariate model adjusted for age, gender, disease duration, treatment, vitamin D levels, BMI and smoking status (smoker or non-smoker). Compared with the baseline intake group (intake below 2 g/day), the average intake (2–4.8 g/day) and above average intake groups (≥4.8 g/day) presented an exacerbation rate of 2.75 (95% CI 1.3 to 5.8) and 3.95-fold higher (95% CI 1.4 to 11.2) than the baseline group (p<0.001 for trend). No significant differences were found in terms of EDSS either at baseline or at the end of follow-up (data not shown).

Sodium intake and MRI outcomes

We then correlated sodium intake levels with radiological activity. As with clinical disease activity, a significant correlation between sodium intake and MRI activity was found including total T2 lesion load and CUA (the combination of new Gd⁺ lesions and new or enlarging T2 lesions; see table 3). Individuals with a sodium intake above the national average had a 3.4-fold increased chance of developing a new lesion on MRI and had, on average, eight more T2 lesions.

Table 3.

Association between sodium intake and radiological activity

	IRR	95% CI	p Value
CUA in MRI (univariate model)
Sodium intake (g/day)
<2	1 (baseline)	–	–
2–4.8	2.68	1.4 to 4.9	0.002
>4.8	3.56	1.7 to 7.55	0.001
CUA in MRI (adjusted model)
Sodium intake (g/day)
<2	1 (baseline)	–	–
2–4.8	2.86	1.52 to 5.4	0.001
>4.8	3.42	1.37 to 8.55	0.008
Age (1-year increment)	0.97	0.93 to 1.00	0.869
Gender (male)	0.57	0.24 to 1.34	0.920
Disease duration (1-year increment)	1.05	0.96 to 1.15	0.25
Vitamin D (1 ng increase)	0.94	0.96 to 1	0.1
Smoking (smoker)	0.51	0.19 to 1.33	0.17
BMI (1 unit increase)	0.92	0.86 to 1	0.06
Treatment (immunosuppressant vs immunomodulators/untreated)	0.45	0.22 to 1.06	0.06
Average T2 lesion count
	Mean	SEM	p Value (vs baseline)
Sodium intake (g/day)
<2	6.45	1.84	–
2–4.8	7.14	0.88	0.743
>4.8	14.13	1.98	0.005

BMI, body mass index; CUA, combined unique activity; IRR, incidence rate ratio.

Correlation between salt intake and MRI activity is also present in group 2

We conducted several analyses to test the robustness of the results such as retesting the same patients at different time points, using a different equation to estimate sodium intake (see online supplementary material) and repeating the analysis in a different subset of patients. We recruited an additional independent cohort to perform a cross-sectional analysis on salt intake and MRI T2 lesion count and CUA to replicate the results found for the original longitudinal cohort. Despite some differences between cohorts (as described previously), there was a positive correlation between salt intake and MS disease severity in the replication group (see table 4).

Table 4.

Association between sodium intake and radiological activity in group 2

	IRR	95% CI	p Value
CUA in MRI (univariate model)
Sodium intake (g/day)
<2	1 (baseline)	–	–
2–4.8	3.9	0.98 to 15.6	0.054
>4.8	1.82	0.4 to 7.7	0.41
CUA in MRI (adjusted model)
Sodium intake (g/day)
<2	1 (baseline)	–	–
2–4.8	3.81	1.1 to 14.6	0.05
>4.8	1.86	0.4 to 7.7	0.39
Gender (male)	0.58	0.2 to 1.9	0.4
Disease duration (1-year increment)	1.1	0.9 to 1.2	0.5
EDSS (1 unit increase)	1.1	0.8 to 1.7	0.5
BMI (1 unit increase)	1.08	0.9 to 1.2	0.4
Treatment (immunosuppressant vs immunomodulators/untreated)	1.1	0.2 to 5.2	0.9
Average T2 lesion count
	Mean	SEM	p Value (vs baseline)
Sodium intake (g/day)
<2	9.43	2.26	–
2–4.8	15.08	1.57	0.055
>4.8	24.44	2.43	<0.001

BMI, body mass index; CUA, combined unique activity; EDSS, Expanded Disability Status Scale; IRR, incidence rate ratio.

Sodium levels in serum are not linked to clinical or radiological outcomes

We then tested whether the aforementioned association observed between sodium intake and clinical and radiological activity extended to sodium serum levels. Owing to its importance in general metabolism, sodium levels are tightly regulated within narrow levels regardless of large variations in dietary sodium consumption.¹⁷ Therefore, overall consumption should not significantly affect serum levels. We did not detect a significant correlation between daily sodium intake and serum sodium after adjusting for BMI, age, gender, vitamin D levels and smoking status (p=0.692), nor there was a correlation between serum sodium and clinical or radiological disease activity (data not shown). Thus, if salt intake has any causal role beyond its association, it does not appear to occur in the peripheral blood.

DISCUSSION

In this study, the risk of clinical or radiological MS exacerbation was increased in (longitudinally followed) individuals with high-sodium intakes. We replicated our results using a cross-sectional analysis of a separate group of 52 patients.

This study has some limitations including a relatively small cohort size and the inability to exclude potential confounders (such as diet, role of commensal microbiota, stress and other conducts that may have an impact on food preference and treatment compliance or healthy overall behavior). Thus, even though an association between increased sodium intake and increased disease activity was shown we cannot claim causality, and we cannot exclude the possibility of reverse causation: individuals with more relapses, received more steroids and thus their salt intake and excretion is increased because they have higher disease activity and not the other way around. Another possible caveat relates to changes in salt consumption over time. We tried to overcome this by retesting the same patients at different time points, finding no significant changes.

Another potential confounder involves patients with increased disease activity who may have hypothalamic lesions and therefore develop inappropriate antidiuretic hormone secretion, as has been reported in patients with MS,¹⁸ and who therefore excrete higher amounts of sodium. However, we ruled out this possibility by measuring the sodium concentration in the serum and by excluding patients with hyponatraemia. There is evidence that sodium intake is linked to obesity and changes in body fat composition,¹⁹ and although obesity has been linked to MS risk, appear to occur in childhood and adolescence.²⁰ Nevertheless, we included this variable in our adjusted models to account for potential confounding factors.

There is widespread evidence from epidemiological and animal studies as well as from clinical trials that dietary salt (sodium chloride) plays a key role in regulating blood pressure.^{11 21} However, the effects of salt consumption on MS may go beyond the effects of elevated blood pressure. Specifically, sodium chloride has been shown to have pleiotropic effects on kidney homeostasis and T-cell function.^22–24 There appears to be a reciprocal relationship between sodium chloride and the immune system: a lack of T and B cells is associated with lower hypertension levels in mice,²⁵ and higher IL-17 levels have been observed in some hypertensive individuals.²⁶ Despite the potential relationship between sodium consumption and the immune system, there is scarce evidence of its role in autoimmune diseases. Twenty years ago a potential relationship between salt and asthma was described,²⁷ and a recent systematic review found evidence that a low sodium diet may improve lung function in exercise induced-asthma, with no clear evidence on the overall control of asthma.²⁸ Other reports have linked sodium intake to mortality in type 1 diabetes, but in this case the involvement of the immune system is less clear and it is probably linked to renal integrity.²⁹

Th17 cells and IL-17 are thought to be major players in MS pathogenesis.^{3 30 31} yet their relationship to sodium chloride levels has been only recently addressed.^{8 9} Increasing sodium chloride concentrations in an in vitro system by 40 mM boosts highly pathogenic human Th17 generation in an serum and glucocorticoid kinase 1-dependent fashion with no significant effects on other effector T-cell subsets.⁹ Furthermore, mice fed a high-salt diet had an exacerbated clinical and histological EAE course that was associated with increased IL-17.^{8 9} Thus, the aforementioned effects of sodium on Th17 cells could at least partly explain the association found in our study.

Future studies should address whether the effects of sodium in MS are mediated by Th17 or additional mechanisms. Sodium chloride intake has many physiological effects such as blood pressure and renin angiotensin system modulation.³² Interestingly, the activation of renin and angiotensin has been implicated in EAE pathogenesis.³³ Moreover, increases in systolic blood pressure similar to those observed with high-salt consumption have recently been shown to be associated with the disruption of white matter integrity in young normotensive individuals.³⁴ Whether high-blood pressure interacts with the typical autoimmune mechanisms associated with MS is an interesting question that remains to be answered.

Nevertheless, our findings suggest that clinical trials with a salt intake reduction as an intervention are needed to establish whether sodium intake control benefits patients with MS.