Dietary sodium and potassium intake in people with diabetes: are guidelines being met?

Cardiovascular diseases are the leading cause of morbidity and mortality, particularly among individuals with diabetes. Blood pressure is a modifiable risk factor, and dietary sodium and potassium play pivotal roles in blood pressure regulation. Public health bodies (e.g., American Diabetes Association, Institute of Medicine) recommend sodium intake below 2300 mg/day (100 mmol/24 h) and potassium intake above 4680 mg/day (120 mmol/24 h). However, global sodium intake typically exceeds recommendations and prior Australian data suggest very high sodium intake among people with diabetes. Evidence linking sodium intake to cardiovascular outcomes is inconsistent, with studies reporting higher, lower, U/J-shaped, uncertain, or no associations. Methodological differences, especially reliance on dietary recall versus 24-h urinary measurements, likely contribute to discrepancies. The present study aims to assess whether people with diabetes adhere to current sodium and potassium guidelines, evaluate cohort-level intake trends from 2009–2015, and quantify individual-level adherence over time. The hypotheses were that adherence would be low and that sodium and potassium intake would remain stable over time.

Prior research shows inconsistent associations between low sodium intake and cardiovascular and all-cause mortality, ranging from higher risk to lower risk, U-/J-shaped relationships, uncertain, or no association. Earlier studies often used dietary recall, which can underestimate sodium intake by about 50%, whereas 24-h urinary sodium excretion provides a more accurate estimate. Australian studies using 24-h urine collections in people with diabetes have shown poor adherence to sodium and potassium recommendations, with only one including both type 1 and type 2 diabetes. Meta-analyses suggest that, in Australia, sodium consumption has not decreased since 1989. Sources of dietary sodium differ by culture; in Western countries, approximately 80% comes from processed foods, limiting the impact of advice to avoid adding salt. Educational interventions (e.g., food label reading) have not consistently reduced 24-h urinary sodium. Potassium intake is globally low, and the sodium-to-potassium ratio may be more predictive of cardiovascular outcomes than either nutrient alone, with lower ratios associated with better outcomes.

Design: Prospective, longitudinal cohort study conducted from 2009 to 2015 at a university teaching hospital diabetes outpatient clinic. Participants: 904 adults with diagnosed diabetes (type 1, type 2, latent autoimmune diabetes in adults, and secondary causes). Written informed consent obtained; ethics approval by Austin Health Human Research Ethics Committee. Measurements: Participants provided 24-h urine collections (total n=3689; average 4 per participant). Urinary measures: sodium (uNa), potassium (uK), volume (uVol), creatinine (uCr), urea (uUrea), glucose (uGlu). Serum measures: creatinine, eGFR (CKD-EPI), urea, uric acid, HbA1c, fasting glucose, total cholesterol, HDL, LDL (calculated via Friedewald equation), triglycerides, C-peptide. Anthropometrics/clinical: age, blood pressure, BMI, duration of diabetes. Laboratory analyzers changed mid-study (Beckman Coulter UniCel DXC800/600 until Jan 2012; Roche Cobas 8000 thereafter). HbA1c by immunoassay on Roche Integra. Definition of adherence: Sodium guideline: uNa < 2300 mg/day (100 mmol/24 h). Potassium guideline: uK > 4680 mg/day (120 mmol/24 h), based on ADA and IOM recommendations for high-risk groups. Data handling and corrections: Because ~90% of ingested sodium and ~80% of potassium are excreted in urine, 24-h excretion was used to estimate intake. To address variation in renal function and potential incomplete collections, uNa values were corrected using urinary creatinine: Corrected uNa = (uNa × overall mean uCr for that individual) / uCr in that sample. No adjustment was made for insensible losses to remain comparable with prior studies. Endpoints: Primary—estimate individuals’ sodium intake over seven years and cohort-level adherence to sodium guidelines. Secondary—estimate potassium intake and adherence at individual and cohort levels. Analysis: Baseline characteristics summarized as mean ± SD. Yearly averages of 24-h uNa and uK were calculated to assess cohort-level adherence and trends. Individual-level adherence calculated as the percentage of a participant’s collections meeting guidelines; summarized as median (IQR). Univariate logistic regression assessed associations between adherence (sodium or potassium) and age, sex, duration of diabetes, HbA1c, fasting glucose, eGFR, serum urea, and lipid profile. Multivariate logistic regression included age, sex, duration of diabetes, and HbA1c. Collinearity tests performed. Statistical significance at p<0.05. Analyses conducted in STATA 14.1.

- Sample: 904 participants (81% type 2 diabetes; 15.9% type 1; 1.1% LADA; others 1.3%); 3689 urine collections (1–16 per participant; average 4). - Mean 24-h excretion (mmol/24 h): sodium 181 ± 73; potassium 76 ± 25. Mean sodium-to-potassium ratio: 2.5 ± 0.9. - Adherence after uCr correction: 7% (63/904) met sodium guideline; 5% (42/904) met potassium guideline. - Medications during collections (n=3689): SGLT2 inhibitors 0.4%; potassium-sparing diuretics 2.2%; loop diuretics 2.2%; thiazide diuretics 18%. - Cohort trends (2009–2015): Mean yearly uNa remained above target (<100 mmol/24 h); mean yearly uK remained below target (>120 mmol/24 h). No change over time for either nutrient at population level. - Individual adherence over time: Median (IQR) percentage of collections meeting guidelines—sodium 0% (25%); potassium 0% (0%). - Univariate logistic regression—sodium adherence: male sex OR 0.40 (p<0.001); duration of diabetes OR 1.04 (p<0.001); age OR 1.01 (p=0.06, NS); eGFR OR 0.99 (p=0.003); triglycerides OR 0.81 (p=0.002); LDL trend OR 0.86 (p=0.05). Other variables NS. - Univariate logistic regression—potassium adherence: male sex OR 6.13 (p<0.001); duration of diabetes OR 0.96 (p=0.006); age OR 0.97 (p=0.007). Other variables NS. - Multivariate logistic regression—sodium adherence: male sex OR 0.38 (p<0.001); duration of diabetes OR 1.04 (p<0.001); HbA1c OR 0.86 (p=0.01); age NS (p=0.54). - Multivariate logistic regression—potassium adherence: male sex OR 5.73 (p<0.001); HbA1c OR 0.77 (p=0.02); duration of diabetes NS (p=0.12); age NS (p=0.21).

The study confirms that people with diabetes rarely meet recommended dietary sodium and potassium targets and that their intake remains stable over time, addressing the hypotheses of low adherence and limited change. At the cohort level, average sodium intake exceeded and potassium intake fell below recommended thresholds in every year studied, and individual-level adherence was persistently low. Sex differences were evident: males were less likely to meet sodium guidelines but more likely to meet potassium guidelines, consistent with greater overall intake. Longer diabetes duration was associated with greater sodium guideline adherence, potentially reflecting health status or dietary modifications over time. The findings align with prior Australian and international research showing persistent high sodium intake and low potassium intake, and they underscore the importance of considering the sodium-to-potassium ratio, which was elevated in this cohort and linked to adverse cardiovascular outcomes in prior literature. Potential contributors include salt appetite, reduced salt taste perception in diabetes, and high reliance on processed foods where most sodium is derived and potassium is depleted. Educational interventions (e.g., label reading) may be insufficient alone to reduce intake. The results, alongside literature suggesting a possible optimal sodium intake range (approximately 115–215 mmol/24 h) and potential U-/J-shaped associations with outcomes, challenge the feasibility and necessity of very stringent sodium targets for all individuals with diabetes. There is a need for pragmatic strategies to reduce sodium in the food supply and increase potassium intake, and for robust randomized trials assessing clinically meaningful outcomes.

People with diabetes in this cohort did not meet dietary sodium and potassium recommendations at either population or individual levels, and intake did not change over time. Adherence was particularly low, with only 7% and 5% meeting sodium and potassium guidelines, respectively. Given the feasibility challenges and mixed observational evidence regarding very low sodium intake, current recommendations may warrant reconsideration. Future research should include long-term interventional and randomized controlled trials to define optimal, achievable sodium and potassium targets and to evaluate the impact on cardiovascular outcomes, as well as strategies to modify the food environment to reduce sodium and increase potassium availability.

- Medications that can affect sodium/potassium excretion (e.g., diuretics, SGLT2 inhibitors) were not excluded, potentially influencing urinary excretion estimates. - Number and timing of urine collections varied between participants, which may introduce variability; however, the average of four collections per participant enhances reliability of habitual intake estimates. - Observational design limits causal inference regarding factors associated with adherence (e.g., HbA1c relationships). - Laboratory analyzer changes occurred mid-study, although standard protocols were followed. - No adjustment for insensible losses (to maintain comparability with prior studies) may slightly underestimate true intake. - Sources of dietary sodium and potassium were not directly assessed, limiting insight into dietary patterns contributing to intake.