To the Editor:
Patients with chronic kidney disease (CKD) are advised to reduce sodium intake, but some studies show higher risks for death and cardiovascular events at lower levels of sodium intake.1, 2, 3 However, many previous studies used a single 24-hour urine collection to estimate sodium intake,4 and studies in CKD were conducted in relatively homogenous European or Australian cohorts.3,5, 6, 7, 8 In contrast, the African American Study of Kidney Disease and Hypertension (AASK) was comprised of self-identified African Americans with significant CKD and ascertained serial 24-hour urine collections.9 We used data from AASK to evaluate the association of sodium excretion with blood pressure (BP), left ventricular mass, and cardiovascular events among African Americans with hypertensive CKD.
For the outcomes of systolic and diastolic BP, we used average values ascertained by 24-hour ambulatory BP monitoring (ABPM) during the cohort phase. Left ventricular mass was measured using 2-dimensional and M-mode imaging from a limited transthoracic echocardiogram during the cohort phase. We examined the cardiovascular composite end point of hospitalization for myocardial infarction, heart failure, stroke, cardiovascular death, or death from any cause.
During the AASK trial phase (1995-1998), 24-hour urine collections were ascertained at baseline and every 6 months thereafter. During the AASK cohort phase (2002-2007), 24-hour urine collections were ascertained at baseline and annually thereafter. We describe baseline characteristics by quintile of mean calibrated 24-hour urinary sodium excretion collected during the first 18 months of the trial phase (median number of urine collections, 3; interquartile range, 2-4). To correct for potential incomplete 24-hour samples, we calibrated urine sodium excretion to mean sex-specific creatinine excretion10: 1,745 mg/24 h for men and 1,194 mg/24 h for women.
We used the %SPECI SAS Macro11 (SAS Institute Inc) to determine the form of sodium excretion that provided the best combination of model fit and clinical applicability for each of our outcomes. To examine the association of sodium excretion with BP and left ventricular mass, we used linear regression, using the mean of all available 24-hour urine collections from randomization up to (but not after) the date of the ABPM or echocardiogram (median number of urine collections, 10; interquartile range, 8-12), respectively. We report the β coefficient for each 1-g increase in urinary sodium excretion and 95% confidence intervals.
To examine the association of sodium excretion with the cardiovascular composite end point, we used proportional hazards regression models and report hazard ratios and 95% confidence intervals. We modeled sodium excretion as a time-varying continuous variable, calculating the cumulative mean urinary sodium excretion using all available 24-hour urine sodium measurements. Follow-up time began at the start of the trial phase. To account for multiple records per patient, we used robust sandwich covariance estimates. Because only 4% of patients were missing 1 or more variables, we performed a complete-case analysis.
All analyses were done using SAS Enterprise Guide 7.15 (SAS Institute Inc). The Institutional Review Board at Stanford University determined that use of the publicly available AASK data set for research did not require further review. This research was conducted under data use agreement 20629 with the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Central Database Repository.
Of the 1,094 participants in AASK, 1,061 had at least one 24-hour urine collection during the first 18 months of the trial phase and were included in the analysis. Overall, mean sodium excretion was 3.7 ± 1.5 g/d (Table 1). Of the 691 participants from the AASK trial phase who enrolled in the cohort phase, 647 had ABPM available. Each 1-g increase in 24-hour sodium excretion was associated with a 1.3/0.9 mm Hg increase in systolic/diastolic BP (Table 2). Higher urinary sodium excretion was associated with higher left ventricular mass (Table 2). During a total of 7,792 person-years of follow-up, there were 354 cardiovascular composite end points (event rate, 4.5/100 person-years). We found no significant association of sodium excretion with the cardiovascular composite end point (Table 2).
Table 1.
Baseline Characteristics by Quintile of 24-Hour Sodium Excretion During the First 18 Months of the AASK Trial Phase
| Q1 | Q2 | Q3 | Q4 | Q5 | P-Trend | |
|---|---|---|---|---|---|---|
| Urine sodium, mg/24 h | 2,079.4 ± 429.3 | 2,886.2 ± 154.5 | 3,430.0 ± 189.3 | 4,160.5 ± 235.9 | 5,937.5 ± 1,345.3 | <0.001 |
| Demographics | ||||||
| Age, y | 54.9 ± 11.0 | 54.7 ± 10.6 | 54.5 ± 11.2 | 53.5 ± 10.8 | 55.0 ± 10.1 | 0.70 |
| Female sex | 131 (61.8%) | 96 (45.3%) | 84 (39.4%) | 56 (26.4%) | 51 (24.1%) | <0.001 |
| Employed | 69 (32.5%) | 84 (39.6%) | 97 (45.5%) | 82 (38.7%) | 70 (33.0%) | >0.99 |
| Lives alone | 50 (23.6%) | 43 (20.3%) | 42 (19.7%) | 43 (20.3%) | 57 (26.9%) | 0.46 |
| <High school education | 80 (37.7%) | 86 (40.6%) | 83 (39.0%) | 88 (41.5%) | 91 (42.9%) | 0.29 |
| <$15,000 annual income | 109 (51.4%) | 96 (45.3%) | 96 (45.1%) | 100 (47.2%) | 104 (49.1%) | 0.79 |
| Insurance status | ||||||
| Private | 87 (41.0%) | 96 (45.3%) | 99 (46.5%) | 82 (38.7%) | 84 (39.6%) | 0.38 |
| Medicare/Medicaid | 52 (24.5%) | 46 (21.7%) | 52 (24.4%) | 51 (24.1%) | 54 (25.5%) | 0.65 |
| None | 73 (34.4%) | 70 (33.0%) | 62 (29.1%) | 79 (37.3%) | 74 (34.9%) | 0.61 |
| Family history of end-stage renal disease | 26 (12.3%) | 22 (10.4%) | 34 (16.0%) | 25 (11.8%) | 31 (14.6%) | 0.40 |
| Left ventricular hypertrophy | 68 (32.1%) | 73 (34.4%) | 77 (36.2%) | 95 (44.8%) | 91 (42.9%) | 0.002 |
| History of heart disease | 108 (50.9%) | 97 (45.8%) | 101 (47.4%) | 116 (54.7%) | 121 (57.1%) | 0.05 |
| Current smoker | 48 (22.6%) | 55 (25.9%) | 62 (29.1%) | 69 (32.5%) | 73 (34.4%) | 0.002 |
| Exercises | 105 (49.5%) | 85 (40.1%) | 104 (48.8%) | 84 (39.6%) | 85 (40.1%) | 0.07 |
| Drinks alcohol | 41 (19.3%) | 49 (23.1%) | 67 (31.5%) | 77 (36.3%) | 61 (28.8%) | 0.001 |
| Measured glomerular filtration rate, mL/min/1.73 m2 | ||||||
| >60 | 38 (17.9%) | 42 (19.8%) | 36 (16.9%) | 39 (18.4%) | 43 (20.3%) | 0.70 |
| 45-59 | 82 (38.7%) | 88 (41.5%) | 81 (38.0%) | 86 (40.6%) | 72 (34.0%) | 0.33 |
| 30-44 | 62 (29.2%) | 50 (23.6%) | 62 (29.1%) | 51 (24.1%) | 61 (28.8%) | 0.96 |
| <30 | 30 (14.2%) | 32 (15.1%) | 34 (16.0%) | 36 (17.0%) | 36 (17.0%) | 0.34 |
| Mean ± SD | 46.6 ± 13.4 | 47.4 ± 13.4 | 46.2 ± 14.0 | 46.9 ± 13.8 | 45.9 ± 13.6 | 0.52 |
| Albumin, g/dL | 4.2 ± 0.3 | 4.2 ± 0.3 | 4.3 ± 0.3 | 4.2 ± 0.3 | 4.2 ± 0.4 | 0.04 |
| Body mass index, kg/m2 | 30.0 ± 6.4 | 30.9 ± 6.7 | 30.5 ± 6.4 | 31.0 ± 6.5 | 30.7 ± 6.8 | 0.32 |
| Systolic blood pressure, mm Hg | 148.5 ± 23.9 | 147.7 ± 24.9 | 148.7 ± 22.8 | 152.2 ± 23.1 | 153.8 ± 24.1 | 0.004 |
| Diastolic blood pressure, mm Hg | 94.0 ± 13.1 | 94.5 ± 15.1 | 94.0 ± 13.7 | 97.3 ± 15.3 | 97.7 ± 13.7 | <0.001 |
| Baseline diuretic use | 113 (53.3%) | 135 (63.7%) | 137 (64.3%) | 128 (60.4%) | 148 (69.8%) | 0.005 |
Note: Categorical variables are given as number (percent); continuous variables are given as mean ± SD. N = 212 for quartiles 1, 2, 4 and 5. N = 213 for quartile 3.
Abbreviation: AASK, African American Study of Kidney Disease and Hypertension; Q, quartile; SD, standard deviation.
Table 2.
Association of 1,000-mg Increase in 24-Hour Sodium Excretion With Specified Outcomes
| N | Unadjusted | Adjusteda | |
|---|---|---|---|
| Systolic blood pressure, mm Hg | 647 | β, 1.23; 95% CI, 0.19 to 2.27 | β, 1.31; 95% CI, 0.21 to 2.42 |
| Diastolic blood pressure, mm Hg | 647 | β, 1.01; 95% CI, 0.36 to 1.67 | β, 0.88; 95% CI, 0.21 to 1.56 |
| Left ventricular mass (2D), g | 613 | β, 18.79; 95% CI, 13.59 to 23.98 | β, 8.81; 95% CI, 3.70 to 13.93 |
| Left ventricular mass (M-mode), g | 621 | β, 14.43; 95% CI, 8.75 to 20.11 | β, 3.31; 95% CI, −2.32 to 8.93 |
| Death or cardiovascular events | 1021 | HR, 1.09; 95% CI, 1.02 to 1.15 | HR, 1.03; 95% CI, 0.96 to 1.11 |
Note: Values shown are the β coefficient or HR as indicated and 95% CIs.
Abbreviations: 2D, 2-dimensional; CI, confidence interval; HR, hazard ratio.
Models were adjusted for age, sex, employment status, income, education, insurance status, family history of end-stage renal disease, exercise, drinks alcohol, current smoking, chronic kidney disease stage, history of heart disease, evidence of left ventricular hypertrophy on electrocardiogram, body mass index, serum albumin level, baseline diuretic use, and randomly assigned blood pressure and antihypertensive medication intervention groups.
In conclusion, using multiple 24-hour urine collections collected in African American patients with hypertensive CKD, we found a modest association between sodium excretion and BP. Although we found an association between sodium excretion and left ventricular mass, we found no significant association of sodium excretion with the cardiovascular composite end point. However, AASK was not originally powered to detect differences in cardiovascular outcomes. Well-designed studies with optimal sodium assessment methodology are needed to determine its relation with cardiovascular and other important clinical outcomes, particularly in patients with CKD.
Article Information
Authors’ Contributions
Research idea and study design: KP, MEM-R, TIC; data acquisition: TIC; data analysis/interpretation: KP, MEM-R, TIC; statistical analysis: MEM-R, TIC; supervision or mentorship: TIC. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved.
Support
None.
Financial Disclosure
Dr Chang reports receiving consulting fees from Janssen, Novo Nordisk, and Fresenius for work unrelated to the current project. The remaining authors declare that they have no relevant financial interests.
Disclaimer
AASK was conducted by the AASK Investigators and supported by the NIDDK. The data from AASK reported here were supplied by the NIDDK Central Repositories. This manuscript was not prepared in collaboration with investigators of AASK and does not necessarily reflect the opinions or views of AASK, the NIDDK Central Repositories, or the NIDDK.
Peer Review
Received July 12, 2019. Evaluated by 1 external peer reviewer, with direct editorial input from the Statistical Editor, an Associate Editor, and the Editor-in-Chief. Accepted in revised form October 18, 2019.
References
- 1.O'Donnell M., Mente A., Rangarajan S. Urinary sodium and potassium excretion, mortality, and cardiovascular events. N Engl J Med. 2014;371:612–623. doi: 10.1056/NEJMoa1311889. [DOI] [PubMed] [Google Scholar]
- 2.Stolarz-Skrzypek K., Kuznetsova T., Thijs L. Fatal and nonfatal outcomes, incidence of hypertension, and blood pressure changes in relation to urinary sodium excretion. JAMA. 2011;305:1777–1785. doi: 10.1001/jama.2011.574. [DOI] [PubMed] [Google Scholar]
- 3.Thomas M.C., Moran J., Forsblom C. The association between dietary sodium intake, ESRD, and all-cause mortality in patients with type 1 diabetes. Diabetes Care. 2011;34:861–866. doi: 10.2337/dc10-1722. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cobb L.K., Anderson C.A.M., Elliott P. Methodological issues in cohort studies that relate sodium intake to cardiovascular disease outcomes: a science advisory from the American Heart Association. Circulation. 2014;129(10):1173–1186. doi: 10.1161/CIR.0000000000000015. [DOI] [PubMed] [Google Scholar]
- 5.Cianciaruso B., Bellizzi V., Minutolo R. Salt intake and renal outcome in patients with progressive renal disease. Miner Electrolyte Metab. 1998;24:296–301. doi: 10.1159/000057385. [DOI] [PubMed] [Google Scholar]
- 6.McMahon E.J., Bauer J.D., Hawley C.M. A randomized trial of dietary sodium restriction in CKD. J Am Soc Nephrol. 2013;24:2096–2103. doi: 10.1681/ASN.2013030285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Smyth A., Dunkler D., Gao P. The relationship between estimated sodium and potassium excretion and subsequent renal outcomes. Kidney Int. 2014;86(6):1205–1212. doi: 10.1038/ki.2014.214. [DOI] [PubMed] [Google Scholar]
- 8.Vegter S., Perna A., Postma M.J., Navis G., Remuzzi G., Ruggenenti P. Sodium intake, ACE inhibition, and progression to ESRD. J Am Soc Nephrol. 2012;23:165–173. doi: 10.1681/ASN.2011040430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wright J.T., Jr., Bakris G., Greene T. Effect of blood pressure lowering and antihypertensive drug class on progression of hypertensive kidney disease: results from the AASK trial. JAMA. 2002;288:2421–2431. doi: 10.1001/jama.288.19.2421. [DOI] [PubMed] [Google Scholar]
- 10.Mills K.T., Chen J., Yang W. Sodium excretion and the risk of cardiovascular disease in patients with chronic kidney disease. JAMA. 2016;315:2200–2210. doi: 10.1001/jama.2016.4447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Liu S, Stedman M. A SAS Macro for Covariate Specification in Linear, Logistic, and Survival Regression: Paper 1223-2017. Proceedings of the SAS Global 2017 Conference. Cary, NC: SAS Institute Inc; 2017.