Global, regional and national sodium intakes in 1990 and 2010: a systematic analysis of 24 h urinary sodium excretion and dietary surveys worldwide

Abstract

Objectives: To estimate global, regional (21 regions) and national (187 countries) sodium intakes in adults in 1990 and 2010.

Design: Bayesian hierarchical modelling using all identifiable primary sources.

Data sources and eligibility: We searched and obtained published and unpublished data from 142 surveys of 24 h urinary sodium and 103 of dietary sodium conducted between 1980 and 2010 across 66 countries. Dietary estimates were converted to urine equivalents based on 79 pairs of dual measurements.

Modelling methods: Bayesian hierarchical modelling used survey data and their characteristics to estimate mean sodium intake, by sex, 5 years age group and associated uncertainty for persons aged 20+ in 187 countries in 1990 and 2010. Country-level covariates were national income/person and composition of food supplies.

Main outcome measures: Mean sodium intake (g/day) as estimable by 24 h urine collections, without adjustment for non-urinary losses.

Results: In 2010, global mean sodium intake was 3.95 g/day (95% uncertainty interval: 3.89 to 4.01). This was nearly twice the WHO recommended limit of 2 g/day and equivalent to 10.06 (9.88-10.21) g/day of salt. Intake in men was ∼10% higher than in women; differences by age were small. Intakes were highest in East Asia, Central Asia and Eastern Europe (mean >4.2 g/day) and in Central Europe and Middle East/North Africa (3.9-4.2 g/day). Regional mean intakes in North America, Western Europe and Australia/New Zealand ranged from 3.4 to 3.8 g/day. Intakes were lower (<3.3 g/day), but more uncertain, in sub-Saharan Africa and Latin America. Between 1990 and 2010, modest, but uncertain, increases in sodium intakes were identified.

Conclusions: Sodium intakes exceed the recommended levels in almost all countries with small differences by age and sex. Virtually all populations would benefit from sodium reduction, supported by enhanced surveillance.

Keywords: Epidemiology; Hypertension < Cardiology; Nutrition & Dietetics; Preventive Medicine.

Figure 1

Search strategies for global sodium exposure data: urine-based (left) and diet-based (right). Note: The black dashed lines show dietary Na assessment studies that were identified through the search for 24 h urinary sodium excretion studies.

Figure 2

Relationship between measured urinary sodium and measured dietary sodium in 79 age/sex strata from 26 survey pairs. Note: The solid line represents the cross-walk regression line (linear regression of log-transformed urinary sodium on log-transformed dietary sodium); and the dotted lines, the 95% uncertainty intervals. With the partial exception of 8 age and sex strata (all from 2 of the 3 InterMap study sites in China) in which urinary sodium levels were systematically higher relative to their dietary levels, good agreement was seen between the two metrics. The regression coefficients obtained from this analysis informed the Bayesian hierarchical model of the relationship between the two metrics. Adding a term to identify the survey sites generating the outlying points increased the R² to 0.79 but it was not retained because this term did not correspond to a generic survey characteristic that could be used in predictions outside the cross-walk dataset. The wider uncertainty bounds for the modelled relationship in the upper part of its range reflect the influence of the outliers.

Figure 3

Sodium intakes in g/day by age, for regions and globally, both sexes combined, 2010.

Figure 4

Mean (95% uncertainty interval) sodium intakes (g/day) in 2010 in 21 Global Burden of Diseases regions. Note: Regions are ranked by levels in both sexes combined, ages 20+. Intakes are not age-standardised.

Figure 5

Mean (95% uncertainty interval) age-standardised sodium intakes (g/day) in 1990 and 2010 in 21 Global Burden of Diseases regions. Note: The upper symbol for each pair is for 2010. Regions are ranked by levels in both sexes combined, ages 20+.

Figure 6

Sodium intakes by country, for ages 20+, average of both sexes, in 2010. (A) Mean intakes in g/day and (B) relative uncertainty*. Note: *Monte Carlo SEs divided by the mean of these intake estimates.