Results From the Trial of Nonpharmacologic Interventions in the Elderly

Am J Public Health. 2016 July; 106(7): 1270–1275.

Liwei Chen, MD, PhD,Zhenzhen Zhang, PhD, Wen Chen, MD, Paul K. Whelton, MD, and Lawrence J. Appel, MD

Liwei Chen is with the Department of Public Health Sciences, Clemson University, Clemson, SC. Zhenzhen Zhang is with the School of Public Health, Oregon Health & Science University, Portland. Wen Chen is with the Department of Pathology, VA Medical Center, Washington, DC. Paul K. Whelton is with the Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA. Lawrence J. Appel is with the Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins Medical Institutions, Baltimore, MD.

Corresponding author.Correspondence should be sent to Liwei Chen, Assistant Professor, Department of Public Health Sciences, Clemson University, 511 Edwards Hall, Clemson, SC 29631 (e-mail: ude. nosmelc@ciewil). Reprints can be ordered at http://www.ajph.org by clicking the “Reprints” link.CONTRIBUTORS

L. Chen wrote the first draft of the article. L. Chen, Z. Zhang, and W. Chen conducted the statistical analyses. P. K. Whelton and L. J. Appel acquired the data. All authors conceptualized and designed the study, interpreted the results, and revised the article.

Peer Reviewed

Accepted February 15, 2016.

Copyright © American Public Health Association 2016This article has been cited by other articles in PMC.

Abstract

Objectives. To determine the effect of sodium (Na) reduction on occurrence of headaches.

Methods. In the Trial of Nonpharmacologic Interventions in the Elderly, 975 men and woman (aged 60–80 years) with hypertension were randomized to a Na-reduction intervention or control group and were followed for up to 36 months. The study was conducted between 1992 and 1995 at 4 clinical centers (Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee).

Results. Mean difference in Na excretion between the Na-reduction intervention and control group was significant at each follow-up visit (P < .001) with an average difference of 38.8 millimoles per 24 hours. The occurrence of headaches was significantly lower in the Na-reduction intervention group (10.5%) compared with control (14.3%) with a hazard ratio of 0.59 (95% confidence interval = 0.40, 0.88; P = .009). The risk of headaches was significantly associated with average level of Na excretion during follow-up, independent of most recent blood pressure. The relationship appeared to be nonlinear with a spline relationship and a knot at 150 millimoles per 24 hours.

Conclusions. Reduced sodium intake, currently recommended for blood pressure control, may also reduce the occurrence of headaches in older persons with hypertension.

Headache is a common health problem in adults, resulting in approximately 18 million physician visits in the United States each year. 1 Globally, the estimated lifetime prevalence of headaches in adults is 66%.2 The most common types of headaches are nonvascular and are commonly termed “tension headaches.”1 Tension headaches have been attributed to muscle spasm in the head, neck, and shoulders in response to stress, fatigue or environmental factors like noise or bright lights.3 However, the pathophysiology of tension headaches is uncertain.

Headaches have been associated with elevated blood pressure (BP), including sustained severe hypertension, malignant hypertension, and paroxysmal hypertension.4 In the Hypertension Optimal Treatment trial, in which participants were randomized to 1 of 3 diastolic BP goals, headaches were reduced in all treatment groups, independent of BP goal and type of antihypertensive drug.5

High sodium intake is associated with elevated BP, and clinical trials have demonstrated that sodium-reduction (NaD) lowers BP in individuals with or without hypertension.6–13 A high intake of sodium potentially leads to headaches through a direct effect on BP or indirectly through a BP-independent mechanism. However, few studies have investigated the relationship of sodium intake to the occurrence of headaches. In preliminary observations from the Trial of Nonpharmacologic Interventions in the Elderly (TONE), we reported that individuals who were assigned to the NaD intervention had a lower incidence of headaches.13 A more recent analysis using data collected from the Dietary Approaches to Stop Hypertension (DASH)-sodium trial reported that NaD was associated with a lower risk of headaches,14 replicating the earlier observation from the TONE study. We aimed to expand on the original observation from TONE by examining the relationship between sodium intake and headaches, with a particular focus on assessing dose–response relationship.

METHODS

TONE was a multicenter, randomized controlled trial designed to test the efficacy of nonpharmacologic interventions as a means to control hypertension in the elderly. A detailed description of the design and methods of this trial has been published elsewhere. 15 Eligible individuals were aged 60 to 80 years, were community dwelling, and had hypertension controlled on single antihypertensive medication. Major exclusion criteria included history of a heart attack or stroke within the preceding 6 months, current angina pectoris, congestive heart failure, insulin-dependent diabetes mellitus, serious mental or physical illness, unexplained or involuntary weight loss of 4.5 kilograms or greater during the previous year, hypercreatinemia (> 2.0 mg/dL), hyperkalemia (> 5.5 mmol/L), and anemia (hemoglobin level < 11 g/dL).

We randomly assigned overweight persons (body mass index [BMI; defined as weight in kilograms divided by the square of height in meters] > 27.3 kg/m² in men, > 27.8 kg/m² in women) to 1 of 4 groups in a 2 × 2 factorial design (usual care, weight loss alone, reduced sodium alone, or combined weight loss and reduced sodium). We assigned nonoverweight persons to usual care or reduced sodium intake. The NaD goal for the reduced sodium groups, both NaD alone and NaD combined with weight loss, was to achieve and maintain a 24-hour dietary sodium intake of 80 millimoles (1800 mg) or less. Three months after the start of intervention, we employed a standardized protocol to gradually taper and withdraw antihypertensive medication in participants whose BP remained less than 150/90 millimeters of mercury.

The primary trial outcome was a composite endpoint that included recurrence of high BP, resumption of antihypertensive medication, or a clinical cardiovascular event. During follow-up, we restarted antihypertensive medication if (1) systolic BP was 190 millimeter of mercury or more or diastolic BP was 110 millimeters of mercury or more at a single visit (average of 3 BP measurements), (2) mean systolic BP was 170 millimeters of mercury or more or diastolic BP was 100 millimeters of mercury or more over 2 consecutive visits (average of 6 BP measurements), or (3) mean systolic BP was 150 millimeters of mercury or more or diastolic BP rose to 90 millimeters of mercury or more at 3 consecutive visits (average of 9 BP measurements).

We enrolled participants between August 30, 1992, and June 27, 1994. We collected TONE data at 2 screening visits to establish study eligibility, a randomization visit, drug withdrawal visits (90 days after the start of intervention), and 11 subsequent quarterly follow-up visits (beginning 6 months after randomization). Closeout visits occurred between July and December 1995. The median follow-up in TONE was 29 months (range = 1–36 months). We conducted the study at 4 clinical centers (Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee).13 Staff members, who were blinded to the participant’s randomized treatment assignment, obtained outcome information.

Variables

We estimated dietary sodium intake by measurement of carefully collected 24-hour urinary sodium excretion. We analyzed urinary sodium levels by flame photometry.13 In TONE we ascertained headaches by participant self-report of an adverse event at any follow-up visit. At each follow-up visit, participants completed a brief self-reported medical history and adverse events questionnaire. A study nurse evaluated any participant who reported an intercurrent health-related concern. If a participant reported a medical problem that was potentially serious (e.g., angina) or a symptom that was severe (e.g., headache or chest pain) the nurse completed an adverse event form. A physician then reviewed and coded the adverse event by type (e.g., stroke, myocardial infarction, headache). We also collected and coded the date of the event. Both the nurse and physician were unaware on the participant’s treatment assignment.

Trained observers who were masked to intervention assignment measured BP. At each visit, we obtained 3 BP measurements while participants rested in the seated position. We collected detailed demographic and socioeconomic information including age, gender, race, physical activity, smoking habits, and alcohol intake at baseline.15 We also collected interval medical information, medication use, and body weight measurement at each visit.

Statistical Analyses

We expressed descriptive data for identification of baseline characteristics as well as urinary sodium excretion at each visit as means (SD) for continuous variables and counts (%) for categorical variables. We used the Student t test and χ² test to compare continuous variables and discrete variables, respectively. For all the analyses, the primary outcome was headache, which we identified by means of adverse event reports.

First, we compared incidence of headaches between those in the NaD intervention group (i.e., the NaD alone plus combined weight loss and NaD interventions) and their counterparts in the control group (i.e., usual care plus weight loss alone interventions). We used Kaplan-Meier plots to explore temporal patterns for cumulative incidence of headaches and Cox proportional hazard regressions to estimate the hazard ratio (HR) and 95% confidence interval (CI) for comparison of first occurrence of a headache between the 2 groups. Because this comparison was on the basis of randomized assignments, we only adjusted for variables that were significantly different between the NaD intervention and control groups (i.e., time-varying drug withdraw status and BPs). Although we conducted the main analyses between the NaD and control groups, we also performed a sensitivity analysis to compare headaches among the 4 original group assignments of the factorial design (i.e., usual care, weight loss alone, NaD alone, and combined weight loss and NaD interventions).

Second, we examined whether there were a dose–response relationship and threshold effect between sodium intake and the occurrence of headaches. The exposure was 24-hour urinary sodium excretion, measured as the average level during follow-up. We used Cox proportional hazard regressions to investigate the frequency of new onset headaches, assessed as an adverse event (incident headaches postrandomization). In these models, we combined participants from all intervention groups and we included potential confounders such as age, gender, race, physical activity, smoking, alcohol intake, headaches at baseline, average weight loss during follow-up, and drug withdraw status (time-varying variable) in the adjusted analyses. To explore whether the relationship of headaches with sodium intake was independent of BP, we additionally adjusted for the most recent BP. We performed all statistical analyses using SAS version 9.4 (SAS Institute, Cary, NC). We considered a P value of less than .05 (2 sides) statistically significant.

RESULTS

displays the baseline characteristics of all TONE study participants by assignment to the NaD intervention or control. Of the 975 individuals, 52% were men, 76% were White, and the mean (SD) age was 65.8 (4.6) years. Mean (SD) 24-hour urinary sodium excretion at baseline was 148.5 (54) millimoles per 24 hours. Overall, there was no statistically significant difference in baseline characteristics by group (NaD vs control).

TABLE 1—

Baseline Characteristics of Participants by Randomly Assigned Group: Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

Characteristic	Control (n = 488), Mean (SD) or %	Sodium Reduction (n = 487), Mean (SD) or %	Alla (n = 975), Mean (SD) or %
Age, y	65. 8 (4.5)	65.8 (4.7)	65.8 (4.6)
Male	51.8	52.6	52.2
White	74.8	77.2	76.0
SBP, mmHg	128.1 (9.3)	128.4 (9.4)	128.2 (9.3)
DBP, mmHg	71.3 (7.4)	71.4 (7.2)	71.4 (7.3)
Urine sodium, mmol/24 h	148. 6 (54.9)	148.4 (53.4)	148.5 (54.2)
Weight, lb	180.0 (26.3)	180.7 (28.4)	180.3 (27.4)
BMI, kg/m2	28.9 (3.5)	28.9 (3.6)	28.9 (3.5)
Alcohol user	34.8	35.5	35.2
Currently smokes	4.9	5.7	5. 3

Participant attendance rates were 90%, 86%, and 86% at the 9-, 18-, and 30-month follow-up visits. BP measurements were available for 100% of the visits. We collected 24-hour urine samples in 867 (88.9%), 804 (82.5%), and 421 (82.1%) participants at the 9-, 18-, and 30-month visits, respectively.

Urinary Sodium Excretion and Headache Occurrence

Baseline urinary sodium excretion did not differ between the NaD and control groups (Table A, available as a supplement to the online version of this article at http://www.ajph.org). We observed significantly greater reductions in urinary sodium from baseline in the NaD group compared with the control group at the 9-, 18-, and 30-month follow-up visits. The mean changes in urinary sodium excretion in the NaD group were −40.9, −43.2, and −45.3 millimoles per 24 hours at the 9-, 18-, and 30-month follow-up visits, respectively. Corresponding mean changes in the control group were 0.4, −3.1, and −4.5 millimoles per 24 hours, respectively. During follow-up, the mean between-group difference in average urinary sodium excretion was 38.8 millimoles per 24 hours.

During follow-up, 126 participants reported headache as an adverse event, and 13 (10.3%) reported headache twice as an adverse event. We used the first reported headache as our study outcome. displays the cumulative incidence of headaches during follow-up by randomized assignment. In the control group, 14.3% reported headaches during follow-up. By contrast, only 10.5% reported headaches in the NaD group (log-rank test, P = .012).

Cumulative Incidence of Headaches by Intervention Group (Reduced Sodium vs Control): Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

Note. We used the log-rank test to compare the group difference (P = .012).

The HR for headaches in the NaD group compared with control was 0.56 (95% CI = 0.38, 0.83; P = .01) after adjustment for clinical center, time-varying drug withdrawal status, and average weight loss during follow-up (model 2). After further adjustment for most recent systolic and diastolic BP before the occurrence of headaches (model 3), the HR was essentially unchanged (HR = 0.59; 95% CI = 0.40, 0.88; P = .02). In the sensitivity analysis, the HR for headaches (model 3) was 0.83 (95% CI = 0.49, 1.40; P = .48) in the weight loos alone intervention, 0.61 (95% CI = 0.39, 0.95; P = .03) in the NaD alone intervention, and 0.47 (95% CI = 0.25, 0.92; P = .03) in the combined weight loos and NaD intervention compared with the control group.

Observational Analyses and Dose–Response Relationship

summarizes results from the Cox proportion hazard models using average urinary sodium excretion during follow-up as the primary exposure. We identified a statistically significant direct association of headaches with average urinary sodium excretion. After adjustment for clinical center, age, smoking, race, gender, alcohol use, physical activity, average weight loss during follow-up, and drug withdrawal status (model 2), a 10 millimoles per 24 hour higher level of urinary sodium excretion was associated with a hazard of headaches that was 7. 0% higher (95% CI = 4%, 11%, P ≤ .001). The association between urinary sodium excretion and risk of headaches persisted after further adjustment for most recent systolic BP and diastolic BP in model 3. Because headaches were more common in women (17.0%) than in men (8.8%), we further stratified our analyses by gender. The magnitude of the associations was similar in both gender; the HR of headaches associated with a 10 millimoles per 24 hours higher level of urinary sodium excretion was 1.09 (95% CI = 1.03, 1.15) for men and 1.07 (95% CI = 1.02, 1.12) for women.

TABLE 2—

HR (95% CI) of Headaches by Level of Urinary Sodium Excretion During Follow-Up: Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

		Average Urinary Sodium as Categorical Variableb
Model	Average Urinary Sodium Excretion as a Continuous Variable,a HR (95% CI)	120–150 mmol/24 h, HR (95% CI)	> 150 mmol/24 h, HR (95% CI)
Model 1c	1. 04 (1.00, 1.07)	1.01 (0.62, 1.66)	1.65 (1.11, 2.46)
Model 2d	1.07 (1.04, 1.11)	1.15 (0.67, 1.95)	2.27 (1.45, 3.55)
Model 3e	1.07 (1.04, 1.11)	1.26 (0.74, 2.16)	2.32 (1.48, 3.65)

In exploratory analyses, the relationship between headaches and absolute urinary sodium level appeared to be nonlinear, with a threshold between 120 to 150 millimoles per 24 hours. We noted the greatest evidence for an association between urinary sodium excretion and occurrence of headaches above this threshold. For this reason, we categorized the participants into 3 groups on the basis of their average urinary sodium excretion during follow-up: (1) less than 120 millimoles per 24 hours (below the threshold, lower sodium intake), (2) 120–150 millimoles per 24 hours (close to threshold, moderate sodium intake), and (3) more than 150 millimoles per 24 hours (above threshold, higher sodium intake). Compared with the category of lower sodium intake, individuals with moderate sodium intake did not exhibit a significantly higher risk of headaches (HR = 1.26; 95% CI = 0.74, 2.16; P = .39). However, individuals in the higher sodium intake category had a significantly higher risk of headaches than did those in the lower sodium intake category (HR = 2.32; 95% CI = 1.48, 3.65; P ≤ .001).

During spline interpolation, we identified knots at 120 millimoles per 24 hours and 150 millimoles per 24 hours. Above the 150 millimoles per 24 hours knot, an increase of 10 millimoles per 24 hours in urinary sodium excretion was associated with an increase of 9.7% in the hazard of headaches (95% CI = 3.4%, 16.9%; P = .002) after adjusting for age, smoking, race, gender, alcohol use, physical activity, average weight loss during follow-up, and drug withdrawal status. This positive association remained significant with further adjustment for most recent BPs in model 3 (HR = 1. 09; 95% CI = 1.02, 1.18; P = .004). The risk of headaches increased dramatically and significantly as the urinary sodium excretion level was above the 150 millimoles per 24 hours (Figure A, available as a supplement to the online version of this article at http://www.ajph.org).

DISCUSSION

In this analysis of 975 persons, aged 60 to 80 years who participated in the TONE trial, we found that NaD was associated with a lower risk of headaches over the course of 36 months of follow-up. Specifically, individuals who were randomly assigned to a NaD intervention had a lower cumulative incidence of headaches than did their counterparts in the control group. In an observational analysis, urinary sodium excretion, measured as an average level during follow-up, was associated with occurrence of incident headaches, independent of other risk factors, baseline headaches, use of BP lowering medications, and most recent level of BP. We observed an apparent threshold effect, with the risk of headaches increasing progressively above a urinary sodium excretion of 150 millimoles per 24 hours. Above the sodium excretion threshold of 150 millimoles per 24 hours, risk of incident headaches was higher by 7.8% for every 10 millimoles average increase in 24-hour urinary sodium; below this threshold, there was no significant relationship.

Although numerous studies have assessed the effects of a reduced sodium intake on BP6,7,11–13 and several studies have assessed the relationship between headaches and BP,4,5,16,17 the relationship between sodium intake and the occurrence of headaches has received scant attention. The most relevant study is the DASH-Sodium trial, a randomized, controlled feeding study that tested the effects of 3 levels of sodium intake (50, 100, and 150 mmol/24 hr) in 2 different diets (the DASH diet and a typical American diet). In a recent secondary analysis of this trial, a reduced sodium intake was associated with significantly lower risk of headache (odds ratio[OR] = 0.69; P = .05, among those consuming a typical American diet, and OR = 0.69; P = . 04, among those consuming the DASH diet) in adults with prehypertension and stage 1 hypertension.14

By contrast to TONE, study participants in the DASH-Sodium trial tended to be younger, with a mean (SD) age of 48 years.10 Also, the 2 trial designs and duration of follow-up were different (a cross-over trial with approximately 30 days of follow-up for each period of intervention in the DASH-Sodium trial and a parallel arm trial with a median follow-up of 29 months in the TONE trial), the extent of NaD was almost twofold greater in the DASH-Sodium trial than in the TONE trial, and the methods used to ascertain headaches were different in the 2 studies (symptom check list in the DASH-Sodium trial and adverse event reporting in the TONE trial). Overall, the findings from our analysis and the corresponding DASH-Sodium report provide consistent evidence regarding the apparent effect of sodium intake on the occurrence of headaches. Because of the use of adverse events reporting as the method of ascertainment for headaches in the TONE trial, it seems likely that the results from this trial may have highlighted the relationship between NaD and more severe headaches.

The mechanisms underlying an association between sodium intake and headaches are uncertain. Sodium intake may be related to headaches through a direct effect on BP or indirectly through other BP-independent mechanisms. Excessive intake of sodium is a well-established risk factor for high BP, and NaD can lower BP in hypertensive and nonhypertensive individuals.6–13 Elevated BP has been associated with headaches independent of weight status and antihypertensive drug therapy.4,5 Therefore, it is possible that a high intake of sodium may lead to headaches through a direct effect on BP. However, controlling for recent BP in the Cox regression models did not attenuate the association between sodium and headaches in our analyses, suggesting that the effect of sodium on headaches might be independent of BP.

One proposed BP-independent mechanism relates to changes in vascular smooth muscle tone. Sodium intake is an important determinant of smooth muscle cell reactivity, which is in part mediated by changes in extracellular volume and perhaps by changes in intracellular calcium. A high sodium intake increases the reactivity of arterioles and the BP response to stress or sympathetic simulation.18 In a study of patients with asthma, a moderate reduction in sodium intake led to an improvement in asthma symptoms, potentially related to changes in bronchial smooth muscle tone.19 Such findings suggest that nonvascular headaches might be part of a continuum with vascular headaches at 1 extreme.20 Still, we cannot totally rule out an effect of BP in our study because we measured BP at prespecified follow-up visits, not at the time of the headaches.

Strengths and Limitations

Our study has several strengths. First, the sample size of TONE trial was relatively large (n = 975), and the duration of follow-up was relatively long (up to 3 years). Second, we collected information on potential adverse events at each study visit; a study clinician who was blinded to the participant’s treatment allocation verified these. Third, we estimated sodium intake by means of carefully collected 24-hour urine samples, with averaging of several samples for the observational analyses: twice before randomization and once at the 9-month, 18-month, and 30-month follow-up visits. Of the available methods, mean 24-hour urinary excretion of sodium collected from multiple occasions provides the most accurate estimate of dietary sodium intake for observational analyses.21 Study limitations include a lack of specific information regarding participant headaches severity and duration and prior history of headaches.

Nevertheless, headaches reported as an adverse event in TONE were associated with baseline headache experience,22 and we adjusted for baseline headaches in our observational analyses. Again, because we did not measured BP at the time of the event, inferences about the relationship between BP and headaches are uncertain. However, there were no differences when we compared BP measured at the study visit closest to the adverse event and the average BP during the follow-ups for study participants. The average time between most recent BP measurement and occurrence of adverse event headache was 12.3 days for those who had adverse event headache, which is relative short. Although the assortment of headaches in this study is probably not subject to reporting bias or directly related to the intervention assignments, we were unable to rule out the possibility that the observed improvement in occurrence of headaches in the NaD group could be a surrogate of improving general well-being.

Conclusions

Our findings suggest that reducing sodium intake lowers the risk of headaches in older persons with hypertension. Our findings are consistent with results from the DASH-Sodium trial, which enrolled a middle-aged study population and used a different approach to ascertainment. Whether a reduced sodium intake lowers the risk of headaches in other populations (e.g., persons with migraine headaches) is unknown. From a policy perspective, these data provide an additional rationale in support of recommendations by the American Heart Association, US federal government, Institute of Medicine, World Health Organization, and others to reduce dietary sodium intake. 23

ACKNOWLEDGMENTS

The Trial of Nonpharmacologic Interventions in the Elderly (TONE) trial was supported by the National Institutes of Health (grants R01 AG-09799, R01 H-48642, R01 AG-09771, R01 AG-09773, P60 AG-10484, and K08 HLO2642).

We thank the TONE participants and staff for their contributions to the study.

HUMAN PARTICIPANT PROTECTION

The institutional review boards at Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee approved the trial protocol. Oversight of the trial was also provided by an external data and safety monitoring board appointed by staff at the National Institute on Aging and the National Heart, Long, and Blood Institute. All potential study participants provided written informed consent. Clinical Trial Registration: clinicaltrials.gov identifier {“type”:”clinical-trial”,”attrs”:{“text”:”NCT00000535″,”term_id”:”NCT00000535″}}NCT00000535.

REFERENCES

1. DuBose CD, Cutlip AC, Cutlip WD., 2nd Migraines and other headaches: an approach to diagnosis and classification. Am Fam Physician. 1995;51(6):1498–1504. 1507–1509. [PubMed] [Google Scholar]2. Stovner L, Hagen K, Jensen R et al. The global burden of headache: a documentation of headache prevalence and disability worldwide. Cephalalgia. 2007;27(3):193–210. [PubMed] [Google Scholar]3. Rasmussen BK, Olesen J. Epidemiology of migraine and tension-type headache. Curr Opin Neurol. 1994;7(3):264–271. [PubMed] [Google Scholar]4. Cirillo M, Stellato D, Lombardi C, De Santo NG, Covelli V. Headache and cardiovascular risk factors: positive association with hypertension. Headache. 1999;39(6):409–416. [PubMed] [Google Scholar]5. Wiklund I, Halling K, Ryden-Bergsten T. [What is the effect of lowering the arterial blood pressure on the quality of life? An auxiliary study to the HOT (Hypertension Optimal Treatment) trial] Arch Mal Coeur Vaiss. 1999;92(8):1079–1082. [PubMed] [Google Scholar]6. The effects of nonpharmacologic interventions on blood pressure of persons with high normal levels. Results of the Trials of Hypertension Prevention, phase I. JAMA. 1992;267(9):1213–1220. [Erratum JAMA 1992;267(17):2330] [PubMed] [Google Scholar]7. Effects of weight loss and sodium reduction intervention on blood pressure and hypertension incidence in overweight people with high-normal blood pressure. The Trials of Hypertension Prevention, phase II. The Trials of Hypertension Prevention Collaborative Research Group. Arch Intern Med. 1997;157(6):657–667. [PubMed] [Google Scholar]8. Aburto NJ, Ziolkovska A, Hooper L, Elliott P, Cappuccio FP, Meerpohl JJ. Effect of lower sodium intake on health: systematic review and meta-analyses. BMJ. 2013;346:f1326. [PMC free article] [PubMed] [Google Scholar]9. Appel LJ American Society of Hypertension Writing Group. ASH position paper: dietary approaches to lower blood pressure. J Clin Hypertens. 2009;11(7):358–368. [PubMed] [Google Scholar]10. Mozaffarian D, Fahimi S, Singh GM et al. Global sodium consumption and death from cardiovascular causes. N Engl J Med. 2014;371(7):624–634. [PubMed] [Google Scholar]11. Sacks FM, Svetkey LP, Vollmer WM et al. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group. N Engl J Med. 2001;344(1):3–10. [PubMed] [Google Scholar]12. Stamler R, Stamler J, Grimm R et al. Nutritional therapy for high blood pressure. Final report of a four-year randomized controlled trial—the Hypertension Control Program. JAMA. 1987;257(11):1484–1491. [PubMed] [Google Scholar]13. Whelton PK, Appel LJ, Espeland MA et al. Sodium reduction and weight loss in the treatment of hypertension in older persons: a randomized controlled trial of nonpharmacologic interventions in the elderly (TONE). TONE Collaborative Research Group. JAMA. 1998;279(11):839–846. [PubMed] [Google Scholar]14. Amer M, Woodward M, Appel LJ. Effects of dietary sodium and the DASH diet on the occurrence of headaches: results from randomised multicentre DASH-sodium clinical trial. BMJ Open. 2014;4(12):e006671. [PMC free article] [PubMed] [Google Scholar]15. Appel LJ, Espeland M, Whelton PK et al. Trial of Nonpharmacologic Intervention in the Elderly (TONE). Design and rationale of a blood pressure control trial. Ann Epidemiol. 1995;5(2):119–129. [PubMed] [Google Scholar]16. Cooper WD, Glover DR, Hormbrey JM, Kimber GR. Headache and blood pressure: evidence of a close relationship. J Hum Hypertens. 1989;3(1):41–44. [PubMed] [Google Scholar]17. Mathew NT. Migraine and hypertension. Cephalalgia. 1999;19(suppl 25):17–19. [PubMed] [Google Scholar]18. Burney P. A diet rich in sodium may potentiate asthma. Epidemiologic evidence for a new hypothesis. Chest. 1987;91(6 suppl):143S–148S. [PubMed] [Google Scholar]19. Carey OJ, Locke C, Cookson JB. Effect of alterations of dietary sodium on the severity of asthma in men. Thorax. 1993;48(7):714–718. [PMC free article] [PubMed] [Google Scholar]20. Merikangas KR, Fenton BT, Cheng SH, Stolar MJ, Risch N. Association between migraine and stroke in a large-scale epidemiological study of the United States. Arch Neurol. 1997;54(4):362–368. [PubMed] [Google Scholar]21. Whelton PK, Appel LJ, Sacco RL et al. Sodium, blood pressure, and cardiovascular disease: further evidence supporting the American Heart Association sodium reduction recommendations. Circulation. 2012;126(24):2880–2889. [PubMed] [Google Scholar]22. Lipton RB, Pfeffer D, Newman LC, Solomon S. Headaches in the elderly. J Pain Symptom Manage. 1993;8(2):87–97. [PubMed] [Google Scholar]23. Whelton PK. Sodium, potassium, blood pressure, and cardiovascular disease in humans. Curr Hypertens Rep. 2014;16(8):465. [PubMed] [Google Scholar]

Results From the Trial of Nonpharmacologic Interventions in the Elderly

Am J Public Health. 2016 July; 106(7): 1270–1275.

Liwei Chen, MD, PhD,Zhenzhen Zhang, PhD, Wen Chen, MD, Paul K. Whelton, MD, and Lawrence J. Appel, MD

Liwei Chen is with the Department of Public Health Sciences, Clemson University, Clemson, SC. Zhenzhen Zhang is with the School of Public Health, Oregon Health & Science University, Portland. Wen Chen is with the Department of Pathology, VA Medical Center, Washington, DC. Paul K. Whelton is with the Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA. Lawrence J. Appel is with the Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins Medical Institutions, Baltimore, MD.

Corresponding author.Correspondence should be sent to Liwei Chen, Assistant Professor, Department of Public Health Sciences, Clemson University, 511 Edwards Hall, Clemson, SC 29631 (e-mail: ude.nosmelc@ciewil). Reprints can be ordered at http://www.ajph.org by clicking the “Reprints” link.CONTRIBUTORS

L. Chen wrote the first draft of the article. L. Chen, Z. Zhang, and W. Chen conducted the statistical analyses. P. K. Whelton and L. J. Appel acquired the data. All authors conceptualized and designed the study, interpreted the results, and revised the article.

Peer Reviewed

Accepted February 15, 2016.

Copyright © American Public Health Association 2016This article has been cited by other articles in PMC.

Abstract

Objectives. To determine the effect of sodium (Na) reduction on occurrence of headaches.

Methods. In the Trial of Nonpharmacologic Interventions in the Elderly, 975 men and woman (aged 60–80 years) with hypertension were randomized to a Na-reduction intervention or control group and were followed for up to 36 months. The study was conducted between 1992 and 1995 at 4 clinical centers (Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee).

Results. Mean difference in Na excretion between the Na-reduction intervention and control group was significant at each follow-up visit (P < .001) with an average difference of 38.8 millimoles per 24 hours. The occurrence of headaches was significantly lower in the Na-reduction intervention group (10. 5%) compared with control (14.3%) with a hazard ratio of 0.59 (95% confidence interval = 0.40, 0.88; P = .009). The risk of headaches was significantly associated with average level of Na excretion during follow-up, independent of most recent blood pressure. The relationship appeared to be nonlinear with a spline relationship and a knot at 150 millimoles per 24 hours.

Conclusions. Reduced sodium intake, currently recommended for blood pressure control, may also reduce the occurrence of headaches in older persons with hypertension.

Headache is a common health problem in adults, resulting in approximately 18 million physician visits in the United States each year.1 Globally, the estimated lifetime prevalence of headaches in adults is 66%.2 The most common types of headaches are nonvascular and are commonly termed “tension headaches.”1 Tension headaches have been attributed to muscle spasm in the head, neck, and shoulders in response to stress, fatigue or environmental factors like noise or bright lights. 3 However, the pathophysiology of tension headaches is uncertain.

Headaches have been associated with elevated blood pressure (BP), including sustained severe hypertension, malignant hypertension, and paroxysmal hypertension.4 In the Hypertension Optimal Treatment trial, in which participants were randomized to 1 of 3 diastolic BP goals, headaches were reduced in all treatment groups, independent of BP goal and type of antihypertensive drug.5

High sodium intake is associated with elevated BP, and clinical trials have demonstrated that sodium-reduction (NaD) lowers BP in individuals with or without hypertension.6–13 A high intake of sodium potentially leads to headaches through a direct effect on BP or indirectly through a BP-independent mechanism. However, few studies have investigated the relationship of sodium intake to the occurrence of headaches. In preliminary observations from the Trial of Nonpharmacologic Interventions in the Elderly (TONE), we reported that individuals who were assigned to the NaD intervention had a lower incidence of headaches. 13 A more recent analysis using data collected from the Dietary Approaches to Stop Hypertension (DASH)-sodium trial reported that NaD was associated with a lower risk of headaches,14 replicating the earlier observation from the TONE study. We aimed to expand on the original observation from TONE by examining the relationship between sodium intake and headaches, with a particular focus on assessing dose–response relationship.

METHODS

TONE was a multicenter, randomized controlled trial designed to test the efficacy of nonpharmacologic interventions as a means to control hypertension in the elderly. A detailed description of the design and methods of this trial has been published elsewhere.15 Eligible individuals were aged 60 to 80 years, were community dwelling, and had hypertension controlled on single antihypertensive medication. Major exclusion criteria included history of a heart attack or stroke within the preceding 6 months, current angina pectoris, congestive heart failure, insulin-dependent diabetes mellitus, serious mental or physical illness, unexplained or involuntary weight loss of 4. 5 kilograms or greater during the previous year, hypercreatinemia (> 2.0 mg/dL), hyperkalemia (> 5.5 mmol/L), and anemia (hemoglobin level < 11 g/dL).

We randomly assigned overweight persons (body mass index [BMI; defined as weight in kilograms divided by the square of height in meters] > 27.3 kg/m² in men, > 27.8 kg/m² in women) to 1 of 4 groups in a 2 × 2 factorial design (usual care, weight loss alone, reduced sodium alone, or combined weight loss and reduced sodium). We assigned nonoverweight persons to usual care or reduced sodium intake. The NaD goal for the reduced sodium groups, both NaD alone and NaD combined with weight loss, was to achieve and maintain a 24-hour dietary sodium intake of 80 millimoles (1800 mg) or less. Three months after the start of intervention, we employed a standardized protocol to gradually taper and withdraw antihypertensive medication in participants whose BP remained less than 150/90 millimeters of mercury.

The primary trial outcome was a composite endpoint that included recurrence of high BP, resumption of antihypertensive medication, or a clinical cardiovascular event. During follow-up, we restarted antihypertensive medication if (1) systolic BP was 190 millimeter of mercury or more or diastolic BP was 110 millimeters of mercury or more at a single visit (average of 3 BP measurements), (2) mean systolic BP was 170 millimeters of mercury or more or diastolic BP was 100 millimeters of mercury or more over 2 consecutive visits (average of 6 BP measurements), or (3) mean systolic BP was 150 millimeters of mercury or more or diastolic BP rose to 90 millimeters of mercury or more at 3 consecutive visits (average of 9 BP measurements).

We enrolled participants between August 30, 1992, and June 27, 1994. We collected TONE data at 2 screening visits to establish study eligibility, a randomization visit, drug withdrawal visits (90 days after the start of intervention), and 11 subsequent quarterly follow-up visits (beginning 6 months after randomization). Closeout visits occurred between July and December 1995. The median follow-up in TONE was 29 months (range = 1–36 months). We conducted the study at 4 clinical centers (Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee).13 Staff members, who were blinded to the participant’s randomized treatment assignment, obtained outcome information.

Variables

We estimated dietary sodium intake by measurement of carefully collected 24-hour urinary sodium excretion. We analyzed urinary sodium levels by flame photometry.13 In TONE we ascertained headaches by participant self-report of an adverse event at any follow-up visit. At each follow-up visit, participants completed a brief self-reported medical history and adverse events questionnaire. A study nurse evaluated any participant who reported an intercurrent health-related concern. If a participant reported a medical problem that was potentially serious (e.g., angina) or a symptom that was severe (e.g., headache or chest pain) the nurse completed an adverse event form. A physician then reviewed and coded the adverse event by type (e. g., stroke, myocardial infarction, headache). We also collected and coded the date of the event. Both the nurse and physician were unaware on the participant’s treatment assignment.

Trained observers who were masked to intervention assignment measured BP. At each visit, we obtained 3 BP measurements while participants rested in the seated position. We collected detailed demographic and socioeconomic information including age, gender, race, physical activity, smoking habits, and alcohol intake at baseline.15 We also collected interval medical information, medication use, and body weight measurement at each visit.

Statistical Analyses

We expressed descriptive data for identification of baseline characteristics as well as urinary sodium excretion at each visit as means (SD) for continuous variables and counts (%) for categorical variables. We used the Student t test and χ² test to compare continuous variables and discrete variables, respectively. For all the analyses, the primary outcome was headache, which we identified by means of adverse event reports.

First, we compared incidence of headaches between those in the NaD intervention group (i.e., the NaD alone plus combined weight loss and NaD interventions) and their counterparts in the control group (i.e., usual care plus weight loss alone interventions). We used Kaplan-Meier plots to explore temporal patterns for cumulative incidence of headaches and Cox proportional hazard regressions to estimate the hazard ratio (HR) and 95% confidence interval (CI) for comparison of first occurrence of a headache between the 2 groups. Because this comparison was on the basis of randomized assignments, we only adjusted for variables that were significantly different between the NaD intervention and control groups (i.e., time-varying drug withdraw status and BPs). Although we conducted the main analyses between the NaD and control groups, we also performed a sensitivity analysis to compare headaches among the 4 original group assignments of the factorial design (i.e., usual care, weight loss alone, NaD alone, and combined weight loss and NaD interventions).

Second, we examined whether there were a dose–response relationship and threshold effect between sodium intake and the occurrence of headaches. The exposure was 24-hour urinary sodium excretion, measured as the average level during follow-up. We used Cox proportional hazard regressions to investigate the frequency of new onset headaches, assessed as an adverse event (incident headaches postrandomization). In these models, we combined participants from all intervention groups and we included potential confounders such as age, gender, race, physical activity, smoking, alcohol intake, headaches at baseline, average weight loss during follow-up, and drug withdraw status (time-varying variable) in the adjusted analyses. To explore whether the relationship of headaches with sodium intake was independent of BP, we additionally adjusted for the most recent BP. We performed all statistical analyses using SAS version 9.4 (SAS Institute, Cary, NC). We considered a P value of less than . 05 (2 sides) statistically significant.

RESULTS

displays the baseline characteristics of all TONE study participants by assignment to the NaD intervention or control. Of the 975 individuals, 52% were men, 76% were White, and the mean (SD) age was 65.8 (4.6) years. Mean (SD) 24-hour urinary sodium excretion at baseline was 148.5 (54) millimoles per 24 hours. Overall, there was no statistically significant difference in baseline characteristics by group (NaD vs control).

TABLE 1—

Baseline Characteristics of Participants by Randomly Assigned Group: Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

Characteristic	Control (n = 488), Mean (SD) or %	Sodium Reduction (n = 487), Mean (SD) or %	Alla (n = 975), Mean (SD) or %
Age, y	65. 8 (4.5)	65.8 (4.7)	65.8 (4.6)
Male	51.8	52.6	52.2
White	74.8	77.2	76.0
SBP, mmHg	128.1 (9.3)	128.4 (9.4)	128.2 (9.3)
DBP, mmHg	71.3 (7.4)	71.4 (7.2)	71.4 (7.3)
Urine sodium, mmol/24 h	148. 6 (54.9)	148.4 (53.4)	148.5 (54.2)
Weight, lb	180.0 (26.3)	180.7 (28.4)	180.3 (27.4)
BMI, kg/m2	28.9 (3.5)	28.9 (3.6)	28.9 (3.5)
Alcohol user	34.8	35.5	35.2
Currently smokes	4.9	5.7	5. 3

Participant attendance rates were 90%, 86%, and 86% at the 9-, 18-, and 30-month follow-up visits. BP measurements were available for 100% of the visits. We collected 24-hour urine samples in 867 (88.9%), 804 (82.5%), and 421 (82.1%) participants at the 9-, 18-, and 30-month visits, respectively.

Urinary Sodium Excretion and Headache Occurrence

Baseline urinary sodium excretion did not differ between the NaD and control groups (Table A, available as a supplement to the online version of this article at http://www.ajph.org). We observed significantly greater reductions in urinary sodium from baseline in the NaD group compared with the control group at the 9-, 18-, and 30-month follow-up visits. The mean changes in urinary sodium excretion in the NaD group were −40.9, −43.2, and −45.3 millimoles per 24 hours at the 9-, 18-, and 30-month follow-up visits, respectively. Corresponding mean changes in the control group were 0.4, −3.1, and −4.5 millimoles per 24 hours, respectively. During follow-up, the mean between-group difference in average urinary sodium excretion was 38.8 millimoles per 24 hours.

During follow-up, 126 participants reported headache as an adverse event, and 13 (10.3%) reported headache twice as an adverse event. We used the first reported headache as our study outcome. displays the cumulative incidence of headaches during follow-up by randomized assignment. In the control group, 14.3% reported headaches during follow-up. By contrast, only 10.5% reported headaches in the NaD group (log-rank test, P = .012).

Cumulative Incidence of Headaches by Intervention Group (Reduced Sodium vs Control): Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

Note. We used the log-rank test to compare the group difference (P = .012).

The HR for headaches in the NaD group compared with control was 0.56 (95% CI = 0.38, 0.83; P = .01) after adjustment for clinical center, time-varying drug withdrawal status, and average weight loss during follow-up (model 2). After further adjustment for most recent systolic and diastolic BP before the occurrence of headaches (model 3), the HR was essentially unchanged (HR = 0.59; 95% CI = 0.40, 0.88; P = .02). In the sensitivity analysis, the HR for headaches (model 3) was 0.83 (95% CI = 0.49, 1.40; P = .48) in the weight loos alone intervention, 0.61 (95% CI = 0.39, 0.95; P = .03) in the NaD alone intervention, and 0.47 (95% CI = 0.25, 0.92; P = .03) in the combined weight loos and NaD intervention compared with the control group.

Observational Analyses and Dose–Response Relationship

summarizes results from the Cox proportion hazard models using average urinary sodium excretion during follow-up as the primary exposure. We identified a statistically significant direct association of headaches with average urinary sodium excretion. After adjustment for clinical center, age, smoking, race, gender, alcohol use, physical activity, average weight loss during follow-up, and drug withdrawal status (model 2), a 10 millimoles per 24 hour higher level of urinary sodium excretion was associated with a hazard of headaches that was 7. 0% higher (95% CI = 4%, 11%, P ≤ .001). The association between urinary sodium excretion and risk of headaches persisted after further adjustment for most recent systolic BP and diastolic BP in model 3. Because headaches were more common in women (17.0%) than in men (8.8%), we further stratified our analyses by gender. The magnitude of the associations was similar in both gender; the HR of headaches associated with a 10 millimoles per 24 hours higher level of urinary sodium excretion was 1.09 (95% CI = 1.03, 1.15) for men and 1.07 (95% CI = 1.02, 1.12) for women.

TABLE 2—

HR (95% CI) of Headaches by Level of Urinary Sodium Excretion During Follow-Up: Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

		Average Urinary Sodium as Categorical Variableb
Model	Average Urinary Sodium Excretion as a Continuous Variable,a HR (95% CI)	120–150 mmol/24 h, HR (95% CI)	> 150 mmol/24 h, HR (95% CI)
Model 1c	1. 04 (1.00, 1.07)	1.01 (0.62, 1.66)	1.65 (1.11, 2.46)
Model 2d	1.07 (1.04, 1.11)	1.15 (0.67, 1.95)	2.27 (1.45, 3.55)
Model 3e	1.07 (1.04, 1.11)	1.26 (0.74, 2.16)	2.32 (1.48, 3.65)

In exploratory analyses, the relationship between headaches and absolute urinary sodium level appeared to be nonlinear, with a threshold between 120 to 150 millimoles per 24 hours. We noted the greatest evidence for an association between urinary sodium excretion and occurrence of headaches above this threshold. For this reason, we categorized the participants into 3 groups on the basis of their average urinary sodium excretion during follow-up: (1) less than 120 millimoles per 24 hours (below the threshold, lower sodium intake), (2) 120–150 millimoles per 24 hours (close to threshold, moderate sodium intake), and (3) more than 150 millimoles per 24 hours (above threshold, higher sodium intake). Compared with the category of lower sodium intake, individuals with moderate sodium intake did not exhibit a significantly higher risk of headaches (HR = 1.26; 95% CI = 0.74, 2.16; P = .39). However, individuals in the higher sodium intake category had a significantly higher risk of headaches than did those in the lower sodium intake category (HR = 2.32; 95% CI = 1.48, 3.65; P ≤ .001).

During spline interpolation, we identified knots at 120 millimoles per 24 hours and 150 millimoles per 24 hours. Above the 150 millimoles per 24 hours knot, an increase of 10 millimoles per 24 hours in urinary sodium excretion was associated with an increase of 9.7% in the hazard of headaches (95% CI = 3.4%, 16.9%; P = .002) after adjusting for age, smoking, race, gender, alcohol use, physical activity, average weight loss during follow-up, and drug withdrawal status. This positive association remained significant with further adjustment for most recent BPs in model 3 (HR = 1. 09; 95% CI = 1.02, 1.18; P = .004). The risk of headaches increased dramatically and significantly as the urinary sodium excretion level was above the 150 millimoles per 24 hours (Figure A, available as a supplement to the online version of this article at http://www.ajph.org).

DISCUSSION

In this analysis of 975 persons, aged 60 to 80 years who participated in the TONE trial, we found that NaD was associated with a lower risk of headaches over the course of 36 months of follow-up. Specifically, individuals who were randomly assigned to a NaD intervention had a lower cumulative incidence of headaches than did their counterparts in the control group. In an observational analysis, urinary sodium excretion, measured as an average level during follow-up, was associated with occurrence of incident headaches, independent of other risk factors, baseline headaches, use of BP lowering medications, and most recent level of BP. We observed an apparent threshold effect, with the risk of headaches increasing progressively above a urinary sodium excretion of 150 millimoles per 24 hours. Above the sodium excretion threshold of 150 millimoles per 24 hours, risk of incident headaches was higher by 7.8% for every 10 millimoles average increase in 24-hour urinary sodium; below this threshold, there was no significant relationship.

Although numerous studies have assessed the effects of a reduced sodium intake on BP6,7,11–13 and several studies have assessed the relationship between headaches and BP,4,5,16,17 the relationship between sodium intake and the occurrence of headaches has received scant attention. The most relevant study is the DASH-Sodium trial, a randomized, controlled feeding study that tested the effects of 3 levels of sodium intake (50, 100, and 150 mmol/24 hr) in 2 different diets (the DASH diet and a typical American diet). In a recent secondary analysis of this trial, a reduced sodium intake was associated with significantly lower risk of headache (odds ratio[OR] = 0.69; P = .05, among those consuming a typical American diet, and OR = 0.69; P = . 04, among those consuming the DASH diet) in adults with prehypertension and stage 1 hypertension.14

By contrast to TONE, study participants in the DASH-Sodium trial tended to be younger, with a mean (SD) age of 48 years.10 Also, the 2 trial designs and duration of follow-up were different (a cross-over trial with approximately 30 days of follow-up for each period of intervention in the DASH-Sodium trial and a parallel arm trial with a median follow-up of 29 months in the TONE trial), the extent of NaD was almost twofold greater in the DASH-Sodium trial than in the TONE trial, and the methods used to ascertain headaches were different in the 2 studies (symptom check list in the DASH-Sodium trial and adverse event reporting in the TONE trial). Overall, the findings from our analysis and the corresponding DASH-Sodium report provide consistent evidence regarding the apparent effect of sodium intake on the occurrence of headaches. Because of the use of adverse events reporting as the method of ascertainment for headaches in the TONE trial, it seems likely that the results from this trial may have highlighted the relationship between NaD and more severe headaches.

The mechanisms underlying an association between sodium intake and headaches are uncertain. Sodium intake may be related to headaches through a direct effect on BP or indirectly through other BP-independent mechanisms. Excessive intake of sodium is a well-established risk factor for high BP, and NaD can lower BP in hypertensive and nonhypertensive individuals.6–13 Elevated BP has been associated with headaches independent of weight status and antihypertensive drug therapy.4,5 Therefore, it is possible that a high intake of sodium may lead to headaches through a direct effect on BP. However, controlling for recent BP in the Cox regression models did not attenuate the association between sodium and headaches in our analyses, suggesting that the effect of sodium on headaches might be independent of BP.

One proposed BP-independent mechanism relates to changes in vascular smooth muscle tone. Sodium intake is an important determinant of smooth muscle cell reactivity, which is in part mediated by changes in extracellular volume and perhaps by changes in intracellular calcium. A high sodium intake increases the reactivity of arterioles and the BP response to stress or sympathetic simulation.18 In a study of patients with asthma, a moderate reduction in sodium intake led to an improvement in asthma symptoms, potentially related to changes in bronchial smooth muscle tone.19 Such findings suggest that nonvascular headaches might be part of a continuum with vascular headaches at 1 extreme.20 Still, we cannot totally rule out an effect of BP in our study because we measured BP at prespecified follow-up visits, not at the time of the headaches.

Strengths and Limitations

Our study has several strengths. First, the sample size of TONE trial was relatively large (n = 975), and the duration of follow-up was relatively long (up to 3 years). Second, we collected information on potential adverse events at each study visit; a study clinician who was blinded to the participant’s treatment allocation verified these. Third, we estimated sodium intake by means of carefully collected 24-hour urine samples, with averaging of several samples for the observational analyses: twice before randomization and once at the 9-month, 18-month, and 30-month follow-up visits. Of the available methods, mean 24-hour urinary excretion of sodium collected from multiple occasions provides the most accurate estimate of dietary sodium intake for observational analyses.21 Study limitations include a lack of specific information regarding participant headaches severity and duration and prior history of headaches.

Nevertheless, headaches reported as an adverse event in TONE were associated with baseline headache experience,22 and we adjusted for baseline headaches in our observational analyses. Again, because we did not measured BP at the time of the event, inferences about the relationship between BP and headaches are uncertain. However, there were no differences when we compared BP measured at the study visit closest to the adverse event and the average BP during the follow-ups for study participants. The average time between most recent BP measurement and occurrence of adverse event headache was 12.3 days for those who had adverse event headache, which is relative short. Although the assortment of headaches in this study is probably not subject to reporting bias or directly related to the intervention assignments, we were unable to rule out the possibility that the observed improvement in occurrence of headaches in the NaD group could be a surrogate of improving general well-being.

Conclusions

Our findings suggest that reducing sodium intake lowers the risk of headaches in older persons with hypertension. Our findings are consistent with results from the DASH-Sodium trial, which enrolled a middle-aged study population and used a different approach to ascertainment. Whether a reduced sodium intake lowers the risk of headaches in other populations (e.g., persons with migraine headaches) is unknown. From a policy perspective, these data provide an additional rationale in support of recommendations by the American Heart Association, US federal government, Institute of Medicine, World Health Organization, and others to reduce dietary sodium intake. 23

ACKNOWLEDGMENTS

The Trial of Nonpharmacologic Interventions in the Elderly (TONE) trial was supported by the National Institutes of Health (grants R01 AG-09799, R01 H-48642, R01 AG-09771, R01 AG-09773, P60 AG-10484, and K08 HLO2642).

We thank the TONE participants and staff for their contributions to the study.

HUMAN PARTICIPANT PROTECTION

The institutional review boards at Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee approved the trial protocol. Oversight of the trial was also provided by an external data and safety monitoring board appointed by staff at the National Institute on Aging and the National Heart, Long, and Blood Institute. All potential study participants provided written informed consent. Clinical Trial Registration: clinicaltrials.gov identifier {“type”:”clinical-trial”,”attrs”:{“text”:”NCT00000535″,”term_id”:”NCT00000535″}}NCT00000535.

REFERENCES

1. DuBose CD, Cutlip AC, Cutlip WD., 2nd Migraines and other headaches: an approach to diagnosis and classification. Am Fam Physician. 1995;51(6):1498–1504. 1507–1509. [PubMed] [Google Scholar]2. Stovner L, Hagen K, Jensen R et al. The global burden of headache: a documentation of headache prevalence and disability worldwide. Cephalalgia. 2007;27(3):193–210. [PubMed] [Google Scholar]3. Rasmussen BK, Olesen J. Epidemiology of migraine and tension-type headache. Curr Opin Neurol. 1994;7(3):264–271. [PubMed] [Google Scholar]4. Cirillo M, Stellato D, Lombardi C, De Santo NG, Covelli V. Headache and cardiovascular risk factors: positive association with hypertension. Headache. 1999;39(6):409–416. [PubMed] [Google Scholar]5. Wiklund I, Halling K, Ryden-Bergsten T. [What is the effect of lowering the arterial blood pressure on the quality of life? An auxiliary study to the HOT (Hypertension Optimal Treatment) trial] Arch Mal Coeur Vaiss. 1999;92(8):1079–1082. [PubMed] [Google Scholar]6. The effects of nonpharmacologic interventions on blood pressure of persons with high normal levels. Results of the Trials of Hypertension Prevention, phase I. JAMA. 1992;267(9):1213–1220. [Erratum JAMA 1992;267(17):2330] [PubMed] [Google Scholar]7. Effects of weight loss and sodium reduction intervention on blood pressure and hypertension incidence in overweight people with high-normal blood pressure. The Trials of Hypertension Prevention, phase II. The Trials of Hypertension Prevention Collaborative Research Group. Arch Intern Med. 1997;157(6):657–667. [PubMed] [Google Scholar]8. Aburto NJ, Ziolkovska A, Hooper L, Elliott P, Cappuccio FP, Meerpohl JJ. Effect of lower sodium intake on health: systematic review and meta-analyses. BMJ. 2013;346:f1326. [PMC free article] [PubMed] [Google Scholar]9. Appel LJ American Society of Hypertension Writing Group. ASH position paper: dietary approaches to lower blood pressure. J Clin Hypertens. 2009;11(7):358–368. [PubMed] [Google Scholar]10. Mozaffarian D, Fahimi S, Singh GM et al. Global sodium consumption and death from cardiovascular causes. N Engl J Med. 2014;371(7):624–634. [PubMed] [Google Scholar]11. Sacks FM, Svetkey LP, Vollmer WM et al. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group. N Engl J Med. 2001;344(1):3–10. [PubMed] [Google Scholar]12. Stamler R, Stamler J, Grimm R et al. Nutritional therapy for high blood pressure. Final report of a four-year randomized controlled trial—the Hypertension Control Program. JAMA. 1987;257(11):1484–1491. [PubMed] [Google Scholar]13. Whelton PK, Appel LJ, Espeland MA et al. Sodium reduction and weight loss in the treatment of hypertension in older persons: a randomized controlled trial of nonpharmacologic interventions in the elderly (TONE). TONE Collaborative Research Group. JAMA. 1998;279(11):839–846. [PubMed] [Google Scholar]14. Amer M, Woodward M, Appel LJ. Effects of dietary sodium and the DASH diet on the occurrence of headaches: results from randomised multicentre DASH-sodium clinical trial. BMJ Open. 2014;4(12):e006671. [PMC free article] [PubMed] [Google Scholar]15. Appel LJ, Espeland M, Whelton PK et al. Trial of Nonpharmacologic Intervention in the Elderly (TONE). Design and rationale of a blood pressure control trial. Ann Epidemiol. 1995;5(2):119–129. [PubMed] [Google Scholar]16. Cooper WD, Glover DR, Hormbrey JM, Kimber GR. Headache and blood pressure: evidence of a close relationship. J Hum Hypertens. 1989;3(1):41–44. [PubMed] [Google Scholar]17. Mathew NT. Migraine and hypertension. Cephalalgia. 1999;19(suppl 25):17–19. [PubMed] [Google Scholar]18. Burney P. A diet rich in sodium may potentiate asthma. Epidemiologic evidence for a new hypothesis. Chest. 1987;91(6 suppl):143S–148S. [PubMed] [Google Scholar]19. Carey OJ, Locke C, Cookson JB. Effect of alterations of dietary sodium on the severity of asthma in men. Thorax. 1993;48(7):714–718. [PMC free article] [PubMed] [Google Scholar]20. Merikangas KR, Fenton BT, Cheng SH, Stolar MJ, Risch N. Association between migraine and stroke in a large-scale epidemiological study of the United States. Arch Neurol. 1997;54(4):362–368. [PubMed] [Google Scholar]21. Whelton PK, Appel LJ, Sacco RL et al. Sodium, blood pressure, and cardiovascular disease: further evidence supporting the American Heart Association sodium reduction recommendations. Circulation. 2012;126(24):2880–2889. [PubMed] [Google Scholar]22. Lipton RB, Pfeffer D, Newman LC, Solomon S. Headaches in the elderly. J Pain Symptom Manage. 1993;8(2):87–97. [PubMed] [Google Scholar]23. Whelton PK. Sodium, potassium, blood pressure, and cardiovascular disease in humans. Curr Hypertens Rep. 2014;16(8):465. [PubMed] [Google Scholar]

Results From the Trial of Nonpharmacologic Interventions in the Elderly

Am J Public Health. 2016 July; 106(7): 1270–1275.

Liwei Chen, MD, PhD,Zhenzhen Zhang, PhD, Wen Chen, MD, Paul K. Whelton, MD, and Lawrence J. Appel, MD

Liwei Chen is with the Department of Public Health Sciences, Clemson University, Clemson, SC. Zhenzhen Zhang is with the School of Public Health, Oregon Health & Science University, Portland. Wen Chen is with the Department of Pathology, VA Medical Center, Washington, DC. Paul K. Whelton is with the Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, LA. Lawrence J. Appel is with the Welch Center for Prevention, Epidemiology and Clinical Research, Johns Hopkins Medical Institutions, Baltimore, MD.

Corresponding author.Correspondence should be sent to Liwei Chen, Assistant Professor, Department of Public Health Sciences, Clemson University, 511 Edwards Hall, Clemson, SC 29631 (e-mail: ude.nosmelc@ciewil). Reprints can be ordered at http://www.ajph.org by clicking the “Reprints” link.CONTRIBUTORS

L. Chen wrote the first draft of the article. L. Chen, Z. Zhang, and W. Chen conducted the statistical analyses. P. K. Whelton and L. J. Appel acquired the data. All authors conceptualized and designed the study, interpreted the results, and revised the article.

Peer Reviewed

Accepted February 15, 2016.

Copyright © American Public Health Association 2016This article has been cited by other articles in PMC.

Abstract

Objectives. To determine the effect of sodium (Na) reduction on occurrence of headaches.

Methods. In the Trial of Nonpharmacologic Interventions in the Elderly, 975 men and woman (aged 60–80 years) with hypertension were randomized to a Na-reduction intervention or control group and were followed for up to 36 months. The study was conducted between 1992 and 1995 at 4 clinical centers (Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee).

Results. Mean difference in Na excretion between the Na-reduction intervention and control group was significant at each follow-up visit (P < .001) with an average difference of 38.8 millimoles per 24 hours. The occurrence of headaches was significantly lower in the Na-reduction intervention group (10. 5%) compared with control (14.3%) with a hazard ratio of 0.59 (95% confidence interval = 0.40, 0.88; P = .009). The risk of headaches was significantly associated with average level of Na excretion during follow-up, independent of most recent blood pressure. The relationship appeared to be nonlinear with a spline relationship and a knot at 150 millimoles per 24 hours.

Conclusions. Reduced sodium intake, currently recommended for blood pressure control, may also reduce the occurrence of headaches in older persons with hypertension.

Headache is a common health problem in adults, resulting in approximately 18 million physician visits in the United States each year.1 Globally, the estimated lifetime prevalence of headaches in adults is 66%.2 The most common types of headaches are nonvascular and are commonly termed “tension headaches.”1 Tension headaches have been attributed to muscle spasm in the head, neck, and shoulders in response to stress, fatigue or environmental factors like noise or bright lights. 3 However, the pathophysiology of tension headaches is uncertain.

Headaches have been associated with elevated blood pressure (BP), including sustained severe hypertension, malignant hypertension, and paroxysmal hypertension.4 In the Hypertension Optimal Treatment trial, in which participants were randomized to 1 of 3 diastolic BP goals, headaches were reduced in all treatment groups, independent of BP goal and type of antihypertensive drug.5

High sodium intake is associated with elevated BP, and clinical trials have demonstrated that sodium-reduction (NaD) lowers BP in individuals with or without hypertension.6–13 A high intake of sodium potentially leads to headaches through a direct effect on BP or indirectly through a BP-independent mechanism. However, few studies have investigated the relationship of sodium intake to the occurrence of headaches. In preliminary observations from the Trial of Nonpharmacologic Interventions in the Elderly (TONE), we reported that individuals who were assigned to the NaD intervention had a lower incidence of headaches. 13 A more recent analysis using data collected from the Dietary Approaches to Stop Hypertension (DASH)-sodium trial reported that NaD was associated with a lower risk of headaches,14 replicating the earlier observation from the TONE study. We aimed to expand on the original observation from TONE by examining the relationship between sodium intake and headaches, with a particular focus on assessing dose–response relationship.

METHODS

TONE was a multicenter, randomized controlled trial designed to test the efficacy of nonpharmacologic interventions as a means to control hypertension in the elderly. A detailed description of the design and methods of this trial has been published elsewhere.15 Eligible individuals were aged 60 to 80 years, were community dwelling, and had hypertension controlled on single antihypertensive medication. Major exclusion criteria included history of a heart attack or stroke within the preceding 6 months, current angina pectoris, congestive heart failure, insulin-dependent diabetes mellitus, serious mental or physical illness, unexplained or involuntary weight loss of 4. 5 kilograms or greater during the previous year, hypercreatinemia (> 2.0 mg/dL), hyperkalemia (> 5.5 mmol/L), and anemia (hemoglobin level < 11 g/dL).

We randomly assigned overweight persons (body mass index [BMI; defined as weight in kilograms divided by the square of height in meters] > 27.3 kg/m² in men, > 27.8 kg/m² in women) to 1 of 4 groups in a 2 × 2 factorial design (usual care, weight loss alone, reduced sodium alone, or combined weight loss and reduced sodium). We assigned nonoverweight persons to usual care or reduced sodium intake. The NaD goal for the reduced sodium groups, both NaD alone and NaD combined with weight loss, was to achieve and maintain a 24-hour dietary sodium intake of 80 millimoles (1800 mg) or less. Three months after the start of intervention, we employed a standardized protocol to gradually taper and withdraw antihypertensive medication in participants whose BP remained less than 150/90 millimeters of mercury.

The primary trial outcome was a composite endpoint that included recurrence of high BP, resumption of antihypertensive medication, or a clinical cardiovascular event. During follow-up, we restarted antihypertensive medication if (1) systolic BP was 190 millimeter of mercury or more or diastolic BP was 110 millimeters of mercury or more at a single visit (average of 3 BP measurements), (2) mean systolic BP was 170 millimeters of mercury or more or diastolic BP was 100 millimeters of mercury or more over 2 consecutive visits (average of 6 BP measurements), or (3) mean systolic BP was 150 millimeters of mercury or more or diastolic BP rose to 90 millimeters of mercury or more at 3 consecutive visits (average of 9 BP measurements).

We enrolled participants between August 30, 1992, and June 27, 1994. We collected TONE data at 2 screening visits to establish study eligibility, a randomization visit, drug withdrawal visits (90 days after the start of intervention), and 11 subsequent quarterly follow-up visits (beginning 6 months after randomization). Closeout visits occurred between July and December 1995. The median follow-up in TONE was 29 months (range = 1–36 months). We conducted the study at 4 clinical centers (Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee).13 Staff members, who were blinded to the participant’s randomized treatment assignment, obtained outcome information.

Variables

We estimated dietary sodium intake by measurement of carefully collected 24-hour urinary sodium excretion. We analyzed urinary sodium levels by flame photometry.13 In TONE we ascertained headaches by participant self-report of an adverse event at any follow-up visit. At each follow-up visit, participants completed a brief self-reported medical history and adverse events questionnaire. A study nurse evaluated any participant who reported an intercurrent health-related concern. If a participant reported a medical problem that was potentially serious (e.g., angina) or a symptom that was severe (e.g., headache or chest pain) the nurse completed an adverse event form. A physician then reviewed and coded the adverse event by type (e. g., stroke, myocardial infarction, headache). We also collected and coded the date of the event. Both the nurse and physician were unaware on the participant’s treatment assignment.

Trained observers who were masked to intervention assignment measured BP. At each visit, we obtained 3 BP measurements while participants rested in the seated position. We collected detailed demographic and socioeconomic information including age, gender, race, physical activity, smoking habits, and alcohol intake at baseline.15 We also collected interval medical information, medication use, and body weight measurement at each visit.

Statistical Analyses

We expressed descriptive data for identification of baseline characteristics as well as urinary sodium excretion at each visit as means (SD) for continuous variables and counts (%) for categorical variables. We used the Student t test and χ² test to compare continuous variables and discrete variables, respectively. For all the analyses, the primary outcome was headache, which we identified by means of adverse event reports.

First, we compared incidence of headaches between those in the NaD intervention group (i.e., the NaD alone plus combined weight loss and NaD interventions) and their counterparts in the control group (i.e., usual care plus weight loss alone interventions). We used Kaplan-Meier plots to explore temporal patterns for cumulative incidence of headaches and Cox proportional hazard regressions to estimate the hazard ratio (HR) and 95% confidence interval (CI) for comparison of first occurrence of a headache between the 2 groups. Because this comparison was on the basis of randomized assignments, we only adjusted for variables that were significantly different between the NaD intervention and control groups (i.e., time-varying drug withdraw status and BPs). Although we conducted the main analyses between the NaD and control groups, we also performed a sensitivity analysis to compare headaches among the 4 original group assignments of the factorial design (i.e., usual care, weight loss alone, NaD alone, and combined weight loss and NaD interventions).

Second, we examined whether there were a dose–response relationship and threshold effect between sodium intake and the occurrence of headaches. The exposure was 24-hour urinary sodium excretion, measured as the average level during follow-up. We used Cox proportional hazard regressions to investigate the frequency of new onset headaches, assessed as an adverse event (incident headaches postrandomization). In these models, we combined participants from all intervention groups and we included potential confounders such as age, gender, race, physical activity, smoking, alcohol intake, headaches at baseline, average weight loss during follow-up, and drug withdraw status (time-varying variable) in the adjusted analyses. To explore whether the relationship of headaches with sodium intake was independent of BP, we additionally adjusted for the most recent BP. We performed all statistical analyses using SAS version 9.4 (SAS Institute, Cary, NC). We considered a P value of less than . 05 (2 sides) statistically significant.

RESULTS

displays the baseline characteristics of all TONE study participants by assignment to the NaD intervention or control. Of the 975 individuals, 52% were men, 76% were White, and the mean (SD) age was 65.8 (4.6) years. Mean (SD) 24-hour urinary sodium excretion at baseline was 148.5 (54) millimoles per 24 hours. Overall, there was no statistically significant difference in baseline characteristics by group (NaD vs control).

TABLE 1—

Baseline Characteristics of Participants by Randomly Assigned Group: Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

Characteristic	Control (n = 488), Mean (SD) or %	Sodium Reduction (n = 487), Mean (SD) or %	Alla (n = 975), Mean (SD) or %
Age, y	65. 8 (4.5)	65.8 (4.7)	65.8 (4.6)
Male	51.8	52.6	52.2
White	74.8	77.2	76.0
SBP, mmHg	128.1 (9.3)	128.4 (9.4)	128.2 (9.3)
DBP, mmHg	71.3 (7.4)	71.4 (7.2)	71.4 (7.3)
Urine sodium, mmol/24 h	148. 6 (54.9)	148.4 (53.4)	148.5 (54.2)
Weight, lb	180.0 (26.3)	180.7 (28.4)	180.3 (27.4)
BMI, kg/m2	28.9 (3.5)	28.9 (3.6)	28.9 (3.5)
Alcohol user	34.8	35.5	35.2
Currently smokes	4.9	5.7	5. 3

Participant attendance rates were 90%, 86%, and 86% at the 9-, 18-, and 30-month follow-up visits. BP measurements were available for 100% of the visits. We collected 24-hour urine samples in 867 (88.9%), 804 (82.5%), and 421 (82.1%) participants at the 9-, 18-, and 30-month visits, respectively.

Urinary Sodium Excretion and Headache Occurrence

Baseline urinary sodium excretion did not differ between the NaD and control groups (Table A, available as a supplement to the online version of this article at http://www.ajph.org). We observed significantly greater reductions in urinary sodium from baseline in the NaD group compared with the control group at the 9-, 18-, and 30-month follow-up visits. The mean changes in urinary sodium excretion in the NaD group were −40.9, −43.2, and −45.3 millimoles per 24 hours at the 9-, 18-, and 30-month follow-up visits, respectively. Corresponding mean changes in the control group were 0.4, −3.1, and −4.5 millimoles per 24 hours, respectively. During follow-up, the mean between-group difference in average urinary sodium excretion was 38.8 millimoles per 24 hours.

During follow-up, 126 participants reported headache as an adverse event, and 13 (10.3%) reported headache twice as an adverse event. We used the first reported headache as our study outcome. displays the cumulative incidence of headaches during follow-up by randomized assignment. In the control group, 14.3% reported headaches during follow-up. By contrast, only 10.5% reported headaches in the NaD group (log-rank test, P = .012).

Cumulative Incidence of Headaches by Intervention Group (Reduced Sodium vs Control): Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

Note. We used the log-rank test to compare the group difference (P = .012).

The HR for headaches in the NaD group compared with control was 0.56 (95% CI = 0.38, 0.83; P = .01) after adjustment for clinical center, time-varying drug withdrawal status, and average weight loss during follow-up (model 2). After further adjustment for most recent systolic and diastolic BP before the occurrence of headaches (model 3), the HR was essentially unchanged (HR = 0.59; 95% CI = 0.40, 0.88; P = .02). In the sensitivity analysis, the HR for headaches (model 3) was 0.83 (95% CI = 0.49, 1.40; P = .48) in the weight loos alone intervention, 0.61 (95% CI = 0.39, 0.95; P = .03) in the NaD alone intervention, and 0.47 (95% CI = 0.25, 0.92; P = .03) in the combined weight loos and NaD intervention compared with the control group.

Observational Analyses and Dose–Response Relationship

summarizes results from the Cox proportion hazard models using average urinary sodium excretion during follow-up as the primary exposure. We identified a statistically significant direct association of headaches with average urinary sodium excretion. After adjustment for clinical center, age, smoking, race, gender, alcohol use, physical activity, average weight loss during follow-up, and drug withdrawal status (model 2), a 10 millimoles per 24 hour higher level of urinary sodium excretion was associated with a hazard of headaches that was 7. 0% higher (95% CI = 4%, 11%, P ≤ .001). The association between urinary sodium excretion and risk of headaches persisted after further adjustment for most recent systolic BP and diastolic BP in model 3. Because headaches were more common in women (17.0%) than in men (8.8%), we further stratified our analyses by gender. The magnitude of the associations was similar in both gender; the HR of headaches associated with a 10 millimoles per 24 hours higher level of urinary sodium excretion was 1.09 (95% CI = 1.03, 1.15) for men and 1.07 (95% CI = 1.02, 1.12) for women.

TABLE 2—

HR (95% CI) of Headaches by Level of Urinary Sodium Excretion During Follow-Up: Trial of Nonpharmacologic Interventions in the Elderly, United States, 1992–1995

		Average Urinary Sodium as Categorical Variableb
Model	Average Urinary Sodium Excretion as a Continuous Variable,a HR (95% CI)	120–150 mmol/24 h, HR (95% CI)	> 150 mmol/24 h, HR (95% CI)
Model 1c	1. 04 (1.00, 1.07)	1.01 (0.62, 1.66)	1.65 (1.11, 2.46)
Model 2d	1.07 (1.04, 1.11)	1.15 (0.67, 1.95)	2.27 (1.45, 3.55)
Model 3e	1.07 (1.04, 1.11)	1.26 (0.74, 2.16)	2.32 (1.48, 3.65)

In exploratory analyses, the relationship between headaches and absolute urinary sodium level appeared to be nonlinear, with a threshold between 120 to 150 millimoles per 24 hours. We noted the greatest evidence for an association between urinary sodium excretion and occurrence of headaches above this threshold. For this reason, we categorized the participants into 3 groups on the basis of their average urinary sodium excretion during follow-up: (1) less than 120 millimoles per 24 hours (below the threshold, lower sodium intake), (2) 120–150 millimoles per 24 hours (close to threshold, moderate sodium intake), and (3) more than 150 millimoles per 24 hours (above threshold, higher sodium intake). Compared with the category of lower sodium intake, individuals with moderate sodium intake did not exhibit a significantly higher risk of headaches (HR = 1.26; 95% CI = 0.74, 2.16; P = .39). However, individuals in the higher sodium intake category had a significantly higher risk of headaches than did those in the lower sodium intake category (HR = 2.32; 95% CI = 1.48, 3.65; P ≤ .001).

During spline interpolation, we identified knots at 120 millimoles per 24 hours and 150 millimoles per 24 hours. Above the 150 millimoles per 24 hours knot, an increase of 10 millimoles per 24 hours in urinary sodium excretion was associated with an increase of 9.7% in the hazard of headaches (95% CI = 3.4%, 16.9%; P = .002) after adjusting for age, smoking, race, gender, alcohol use, physical activity, average weight loss during follow-up, and drug withdrawal status. This positive association remained significant with further adjustment for most recent BPs in model 3 (HR = 1. 09; 95% CI = 1.02, 1.18; P = .004). The risk of headaches increased dramatically and significantly as the urinary sodium excretion level was above the 150 millimoles per 24 hours (Figure A, available as a supplement to the online version of this article at http://www.ajph.org).

DISCUSSION

In this analysis of 975 persons, aged 60 to 80 years who participated in the TONE trial, we found that NaD was associated with a lower risk of headaches over the course of 36 months of follow-up. Specifically, individuals who were randomly assigned to a NaD intervention had a lower cumulative incidence of headaches than did their counterparts in the control group. In an observational analysis, urinary sodium excretion, measured as an average level during follow-up, was associated with occurrence of incident headaches, independent of other risk factors, baseline headaches, use of BP lowering medications, and most recent level of BP. We observed an apparent threshold effect, with the risk of headaches increasing progressively above a urinary sodium excretion of 150 millimoles per 24 hours. Above the sodium excretion threshold of 150 millimoles per 24 hours, risk of incident headaches was higher by 7.8% for every 10 millimoles average increase in 24-hour urinary sodium; below this threshold, there was no significant relationship.

Although numerous studies have assessed the effects of a reduced sodium intake on BP6,7,11–13 and several studies have assessed the relationship between headaches and BP,4,5,16,17 the relationship between sodium intake and the occurrence of headaches has received scant attention. The most relevant study is the DASH-Sodium trial, a randomized, controlled feeding study that tested the effects of 3 levels of sodium intake (50, 100, and 150 mmol/24 hr) in 2 different diets (the DASH diet and a typical American diet). In a recent secondary analysis of this trial, a reduced sodium intake was associated with significantly lower risk of headache (odds ratio[OR] = 0.69; P = .05, among those consuming a typical American diet, and OR = 0.69; P = . 04, among those consuming the DASH diet) in adults with prehypertension and stage 1 hypertension.14

By contrast to TONE, study participants in the DASH-Sodium trial tended to be younger, with a mean (SD) age of 48 years.10 Also, the 2 trial designs and duration of follow-up were different (a cross-over trial with approximately 30 days of follow-up for each period of intervention in the DASH-Sodium trial and a parallel arm trial with a median follow-up of 29 months in the TONE trial), the extent of NaD was almost twofold greater in the DASH-Sodium trial than in the TONE trial, and the methods used to ascertain headaches were different in the 2 studies (symptom check list in the DASH-Sodium trial and adverse event reporting in the TONE trial). Overall, the findings from our analysis and the corresponding DASH-Sodium report provide consistent evidence regarding the apparent effect of sodium intake on the occurrence of headaches. Because of the use of adverse events reporting as the method of ascertainment for headaches in the TONE trial, it seems likely that the results from this trial may have highlighted the relationship between NaD and more severe headaches.

The mechanisms underlying an association between sodium intake and headaches are uncertain. Sodium intake may be related to headaches through a direct effect on BP or indirectly through other BP-independent mechanisms. Excessive intake of sodium is a well-established risk factor for high BP, and NaD can lower BP in hypertensive and nonhypertensive individuals.6–13 Elevated BP has been associated with headaches independent of weight status and antihypertensive drug therapy.4,5 Therefore, it is possible that a high intake of sodium may lead to headaches through a direct effect on BP. However, controlling for recent BP in the Cox regression models did not attenuate the association between sodium and headaches in our analyses, suggesting that the effect of sodium on headaches might be independent of BP.

One proposed BP-independent mechanism relates to changes in vascular smooth muscle tone. Sodium intake is an important determinant of smooth muscle cell reactivity, which is in part mediated by changes in extracellular volume and perhaps by changes in intracellular calcium. A high sodium intake increases the reactivity of arterioles and the BP response to stress or sympathetic simulation.18 In a study of patients with asthma, a moderate reduction in sodium intake led to an improvement in asthma symptoms, potentially related to changes in bronchial smooth muscle tone.19 Such findings suggest that nonvascular headaches might be part of a continuum with vascular headaches at 1 extreme.20 Still, we cannot totally rule out an effect of BP in our study because we measured BP at prespecified follow-up visits, not at the time of the headaches.

Strengths and Limitations

Our study has several strengths. First, the sample size of TONE trial was relatively large (n = 975), and the duration of follow-up was relatively long (up to 3 years). Second, we collected information on potential adverse events at each study visit; a study clinician who was blinded to the participant’s treatment allocation verified these. Third, we estimated sodium intake by means of carefully collected 24-hour urine samples, with averaging of several samples for the observational analyses: twice before randomization and once at the 9-month, 18-month, and 30-month follow-up visits. Of the available methods, mean 24-hour urinary excretion of sodium collected from multiple occasions provides the most accurate estimate of dietary sodium intake for observational analyses.21 Study limitations include a lack of specific information regarding participant headaches severity and duration and prior history of headaches.

Nevertheless, headaches reported as an adverse event in TONE were associated with baseline headache experience,22 and we adjusted for baseline headaches in our observational analyses. Again, because we did not measured BP at the time of the event, inferences about the relationship between BP and headaches are uncertain. However, there were no differences when we compared BP measured at the study visit closest to the adverse event and the average BP during the follow-ups for study participants. The average time between most recent BP measurement and occurrence of adverse event headache was 12.3 days for those who had adverse event headache, which is relative short. Although the assortment of headaches in this study is probably not subject to reporting bias or directly related to the intervention assignments, we were unable to rule out the possibility that the observed improvement in occurrence of headaches in the NaD group could be a surrogate of improving general well-being.

Conclusions

Our findings suggest that reducing sodium intake lowers the risk of headaches in older persons with hypertension. Our findings are consistent with results from the DASH-Sodium trial, which enrolled a middle-aged study population and used a different approach to ascertainment. Whether a reduced sodium intake lowers the risk of headaches in other populations (e.g., persons with migraine headaches) is unknown. From a policy perspective, these data provide an additional rationale in support of recommendations by the American Heart Association, US federal government, Institute of Medicine, World Health Organization, and others to reduce dietary sodium intake.23

ACKNOWLEDGMENTS

The Trial of Nonpharmacologic Interventions in the Elderly (TONE) trial was supported by the National Institutes of Health (grants R01 AG-09799, R01 H-48642, R01 AG-09771, R01 AG-09773, P60 AG-10484, and K08 HLO2642).

We thank the TONE participants and staff for their contributions to the study.

HUMAN PARTICIPANT PROTECTION

The institutional review boards at Johns Hopkins University, Wake Forest University School of Medicine, Robert Wood Johnson Medical School, and the University of Tennessee approved the trial protocol. Oversight of the trial was also provided by an external data and safety monitoring board appointed by staff at the National Institute on Aging and the National Heart, Long, and Blood Institute. All potential study participants provided written informed consent. Clinical Trial Registration: clinicaltrials.gov identifier {“type”:”clinical-trial”,”attrs”:{“text”:”NCT00000535″,”term_id”:”NCT00000535″}}NCT00000535.

REFERENCES

1. DuBose CD, Cutlip AC, Cutlip WD., 2nd Migraines and other headaches: an approach to diagnosis and classification. Am Fam Physician. 1995;51(6):1498–1504. 1507–1509. [PubMed] [Google Scholar]2. Stovner L, Hagen K, Jensen R et al. The global burden of headache: a documentation of headache prevalence and disability worldwide. Cephalalgia. 2007;27(3):193–210. [PubMed] [Google Scholar]3. Rasmussen BK, Olesen J. Epidemiology of migraine and tension-type headache. Curr Opin Neurol. 1994;7(3):264–271. [PubMed] [Google Scholar]4. Cirillo M, Stellato D, Lombardi C, De Santo NG, Covelli V. Headache and cardiovascular risk factors: positive association with hypertension. Headache. 1999;39(6):409–416. [PubMed] [Google Scholar]5. Wiklund I, Halling K, Ryden-Bergsten T. [What is the effect of lowering the arterial blood pressure on the quality of life? An auxiliary study to the HOT (Hypertension Optimal Treatment) trial] Arch Mal Coeur Vaiss. 1999;92(8):1079–1082. [PubMed] [Google Scholar]6. The effects of nonpharmacologic interventions on blood pressure of persons with high normal levels. Results of the Trials of Hypertension Prevention, phase I. JAMA. 1992;267(9):1213–1220. [Erratum JAMA 1992;267(17):2330] [PubMed] [Google Scholar]7. Effects of weight loss and sodium reduction intervention on blood pressure and hypertension incidence in overweight people with high-normal blood pressure. The Trials of Hypertension Prevention, phase II. The Trials of Hypertension Prevention Collaborative Research Group. Arch Intern Med. 1997;157(6):657–667. [PubMed] [Google Scholar]8. Aburto NJ, Ziolkovska A, Hooper L, Elliott P, Cappuccio FP, Meerpohl JJ. Effect of lower sodium intake on health: systematic review and meta-analyses. BMJ. 2013;346:f1326. [PMC free article] [PubMed] [Google Scholar]9. Appel LJ American Society of Hypertension Writing Group. ASH position paper: dietary approaches to lower blood pressure. J Clin Hypertens. 2009;11(7):358–368. [PubMed] [Google Scholar]10. Mozaffarian D, Fahimi S, Singh GM et al. Global sodium consumption and death from cardiovascular causes. N Engl J Med. 2014;371(7):624–634. [PubMed] [Google Scholar]11. Sacks FM, Svetkey LP, Vollmer WM et al. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group. N Engl J Med. 2001;344(1):3–10. [PubMed] [Google Scholar]12. Stamler R, Stamler J, Grimm R et al. Nutritional therapy for high blood pressure. Final report of a four-year randomized controlled trial—the Hypertension Control Program. JAMA. 1987;257(11):1484–1491. [PubMed] [Google Scholar]13. Whelton PK, Appel LJ, Espeland MA et al. Sodium reduction and weight loss in the treatment of hypertension in older persons: a randomized controlled trial of nonpharmacologic interventions in the elderly (TONE). TONE Collaborative Research Group. JAMA. 1998;279(11):839–846. [PubMed] [Google Scholar]14. Amer M, Woodward M, Appel LJ. Effects of dietary sodium and the DASH diet on the occurrence of headaches: results from randomised multicentre DASH-sodium clinical trial. BMJ Open. 2014;4(12):e006671. [PMC free article] [PubMed] [Google Scholar]15. Appel LJ, Espeland M, Whelton PK et al. Trial of Nonpharmacologic Intervention in the Elderly (TONE). Design and rationale of a blood pressure control trial. Ann Epidemiol. 1995;5(2):119–129. [PubMed] [Google Scholar]16. Cooper WD, Glover DR, Hormbrey JM, Kimber GR. Headache and blood pressure: evidence of a close relationship. J Hum Hypertens. 1989;3(1):41–44. [PubMed] [Google Scholar]17. Mathew NT. Migraine and hypertension. Cephalalgia. 1999;19(suppl 25):17–19. [PubMed] [Google Scholar]18. Burney P. A diet rich in sodium may potentiate asthma. Epidemiologic evidence for a new hypothesis. Chest. 1987;91(6 suppl):143S–148S. [PubMed] [Google Scholar]19. Carey OJ, Locke C, Cookson JB. Effect of alterations of dietary sodium on the severity of asthma in men. Thorax. 1993;48(7):714–718. [PMC free article] [PubMed] [Google Scholar]20. Merikangas KR, Fenton BT, Cheng SH, Stolar MJ, Risch N. Association between migraine and stroke in a large-scale epidemiological study of the United States. Arch Neurol. 1997;54(4):362–368. [PubMed] [Google Scholar]21. Whelton PK, Appel LJ, Sacco RL et al. Sodium, blood pressure, and cardiovascular disease: further evidence supporting the American Heart Association sodium reduction recommendations. Circulation. 2012;126(24):2880–2889. [PubMed] [Google Scholar]22. Lipton RB, Pfeffer D, Newman LC, Solomon S. Headaches in the elderly. J Pain Symptom Manage. 1993;8(2):87–97. [PubMed] [Google Scholar]23. Whelton PK. Sodium, potassium, blood pressure, and cardiovascular disease in humans. Curr Hypertens Rep. 2014;16(8):465. [PubMed] [Google Scholar]

Does eating more salt prevent migraines and severe headaches?

By Jessica Hamzelou

A bit of salt might stop headaches’ battery of the brain

Chris Fertnig/Getty

Could a salty diet keep migraines at bay? People who eat a lot of salt report having fewer migraines and severe headaches – the first evidence that dietary sodium may affect the condition. But the researchers caution that more evidence is needed before people change their diets, given that high salt consumption is linked to heart disease and stroke.

There is growing evidence linking migraines with sodium. During a migraine, levels of sodium have been found to rise in cerebrospinal fluid, the liquid that bathes the brain and central nervous system. And sodium levels in this liquid seem to peak in the early morning and late afternoon – times of day when people commonly report experiencing migraines.

Plenty of sodium gets into our bodies via the food we eat. “I started to wonder if migraines could be affected by diet,” says Michael Harrington at Huntington Medical Research Institutes in Pasadena, California. To find out, he and his colleagues turned to the National Health and Nutritional Examination Survey, a US survey of the health and diets of tens of thousands of people. Among other things, the survey asks respondents to list everything they consume over a 24-hour period, and whether they experienced a severe headache or migraine during that time.

Of 8819 adults surveyed between 1999 and 2004, the team found that those with the highest levels of sodium in their diets – in products like meat, cheese and bread as well as table salt – reported the fewest severe headaches and migraines.

Harrington says he’s surprised by the results as they are counterintuitive. Given that sodium ions are known to activate neurons, we might have expected the relationship to go in the other direction. High sodium levels generally make neurons more excitable, so the idea that they in some way inhibit or prevent migraine activity is puzzling.

“I think people with migraine handle sodium differently,” says Harrington.

American diets

The theory makes sense, says Svetlana Blitshteyn, who treats and studies nervous system disorders at the University at Buffalo School of Medicine and Biomedical Sciences in New York.

Blitshteyn specialises in disorders of the autonomic nervous system, which controls all our automatic functions, such as heart rate, breathing and urination. Many of her patients have migraines too, and she has noticed that, if they start consuming more salt as a treatment for a different condition, their migraine symptoms often get better – although this evidence is only anecdotal and hasn’t been published yet.

But it is too early to know how safe eating more salt is for people who have migraines, and who might benefit from doing so. “We need more evidence before we can make general recommendations,” says Blitshteyn.

Harrington agrees. Salt has its own risks, and has been linked to heart disease and stroke. Harrington points out that almost all of the people surveyed in his study were on typical US diets, which are already high in salt. Until we know more, the best advice for people with migraines is to eat well and regularly.

Journal reference: Headache, DOI: 10.1111/head.12792

More on these topics:

Cut Back on Salt to Decrease Headaches • National Headache Foundation

15 Jan Cut Back on Salt to Decrease Headaches

Posted at 23:02h
in News to Know
by headache

Individuals who struggle with headaches may want to bypass the salt shaker.

A study published last month in the journal BMJ Open shows diets high in salty foods are a major culprit in headaches and that reducing salt brings relief.

Nearly 400 people participated in the study, and those who ate foods high in sodium (8 grams a day, or about 1.5 tsp.) had nearly one-third more headaches than people who ate foods low in sodium (about 4 grams a day). The highest occurrence of headache was reported by participants eating a typical North American with high sodium (47%), and the lowest occurrence was reported by participants on the DASH diet with low sodium (36%). The DASH diet, which many know as a diet to help control blood pressure, is low in fat and rich in fruits and vegetables.

Regardless of which diets the participants followed, researchers found that those who ate the most sodium reported the highest frequency of headaches, and those who ate the least sodium reported the least amount of headaches.

“A reduced sodium intake was associated with a significantly lower risk of headache, while dietary patterns had no effect on the risk of headaches in adults,” said Lawrence Appel, MD, MPH, of John Hopkins University. “Reduced dietary sodium intake offers a novel approach to prevent headaches.”

While high blood pressure is often associated with high salt intake and headache, the researchers found that people with normal blood pressure had fewer headaches when they reduced salt intake. They noted that a process that is independent of blood pressure may mediate the relationship between sodium and headaches.

The researchers added that the data on the relationship between sodium intake and any form of headaches are sparse, and more study is needed to replicate their findings and further explore the link between sodium intake and headache.

What Is the Verdict on Migraine and Salt?

A closer look at Meniere’s Disease, Migraine and Salt

Excerpt from The Migraine Relief Plan: An 8-Week Transition to Better Eating, Fewer Headaches, and Optimal Health

The Link Between Heart Disease, Migraine and Salt

Those who live with migraine or Meniere’s disease need to be particularly cautious when it comes to sodium. For one thing, a high-sodium meal may trigger headaches.

Sodium and Migraine: The Latest Research

Editor’s Note: This section is NOT an excerpt from The Migraine Relief Plan.

Researchers have known for a while that sodium levels within the brain increase during a migraine attack. New research led by Michael Harrington from the Huntington Medical Research Insitute and Samuel Grant from the FAMU-FSU College of Engineering with Linda Petzold and Hamed Ghaffari from the University of California, Santa Barbara, sheds light on the salt and migraine connection.

The team hypothesized that sodium enters the cerebrospinal fluid (the fluid that surrounds the brain and spinal cord) by passing through the blood-brain barrier (BBB) and the blood-CSF barrier (BCSFB), which separates CSF from blood and the brain.

The researchers used a model of a rat brain to watch how sodium moves between CSF, blood, and the brain during a migraine attack.

“Cerebrospinal fluid (CSF) and brain tissue sodium levels increase during migraine, ” said Linda Petzold, Professor at UCSB in the Department of Mechanical Engineering and the Department of Computer Science. “However, little has been known regarding the underlying mechanisms of sodium homeostasis disturbance in the brain during the onset and propagation of migraine. Using mathematical modeling and global sensitivity analysis, we established that sodium transport from blood to CSF across the blood-CSF barrier is the most likely source of the increased sodium levels.”

It is not completely clear from this research what the role of sodium in your diet would play in your migraine attacks, but it does answer some interesting questions about what happens in your brain during an attack.

Tips for Healthy Salt Consumption

Today, the American Heart Association recommends no more than 1,500 milligrams per day be consumed by everyone.

My recommendation is to live between 1,200 and 1,500 milligrams per day, although if you have Meniere’s disease or any balance problems, experiment with maintaining a consistent average on the lower end of that range. Here are a couple of tips to keep in mind:

1 – Be patient. Once you reduce your sodium intake, it takes two months, possibly three, for your palate to heal and get used to low-sodium foods. They do eventually taste great—I promise!

2 – Don’t eliminate all salt. If you eliminate processed foods and salt from your cooking, you can enjoy adding sea salt to your food at the table where you really taste it.

4 – Don’t try to cure migraine with salt. You may have seen on the Internet that taking a large quantity of salt will instantly “cure” a migraine. I don’t recommend that approach, nor does any reputable doctor.

Keep in mind that low-sodium diets can improve more than your migraine. They can also improve:

• Diabetes symptoms
• Carpal tunnel syndrome
• PMS symptoms
• Joint issues and joint pain
• Meniere’s Disease-related dizziness and vertigo (It’s important to maintain a daily average that’s consistent, ideally around 1,000 milligrams per day.)

Don’t miss Dr. Cynthia Armand discuss food and migraine on the Migraine World Summit – Watch here

Salt-free flavor intensifiers

When you start eating lower-sodium foods, it takes a while to adjust, and food doesn’t have that zing. I found that these foods can help add that punch I used to get from salt. As you clean out your pantry, think about items you’ll need to replace, like onion salt, and look for an alternative to try.

• Diced celery, frozen in one-cup servings
• Thinly sliced leeks, white and light green parts only, frozen in one-cup servings
• Roasted garlic, frozen in ice cube trays, then stored in a zip-top freezer bag
• Roasted red peppers, puréed and frozen in ice cube trays (You can use jarred peppers if they have no added salt, vinegar, or sodium bisulfite.)
• Oven-roasted tomatoes drizzled with olive oil, garlic, and black pepper
• Mushroom broth, reduce down (especially shiitake)
• Mushroom purée (cooked)
• Unsalted chicken broth reduced down
• Tomato paste (no salt added)
• Truffle oil
• Infused olive oil, such as basil, garlic, chili pepper
• Roasted eggplant purée

Remember that tomatoes and mushrooms are naturally high in glutamates, which may be a migraine trigger for some people.

Reprinted with permission from The Migraine Relief Plan by Stephanie Weaver, Agate Surrey, 2017.

Updated March 2020

Effects of dietary sodium and the DASH diet on the occurrence of headaches: results from randomised multicentre DASH-Sodium clinical trial

Keywords

Strengths and limitations of this study

• Post hoc analysis of a multicentre randomised clinical trial comparing effects of the two diet patterns using parallel design together with a three-period crossover of three levels of dietary sodium (high (150 mmol), intermediate (100 mmol), low (50 mmol)) on headaches in healthy adults with stage 1 hypertension.

• Three screening and two run-in feeding periods prior to randomisation to assess participant’s eligibility, compliance with dietary requirements and to estimate caloric requirements to maintain weight during study.

• Vigorous efforts made to promote adherence with assigned diets during feeding periods.

• Lack of information on the prevalence of headaches at baseline as well as type of self-reported headaches experienced by participants at the end of feeding periods.

Introduction

Worldwide, headache is a common medical problem and among the most frequently reported disorders of the nervous system.1–3 Globally, 46% of adults are estimated to have an active headache disorder (42% for tension-type headaches; 11% for migraines).2 ,4–6 Headaches affect all age groups, with a higher prevalence in women compared with men.4–6 The direct cost of healthcare services, and medications for the management of headaches is likewise substantial,7–11 as are indirect costs. Patients with frequent headaches have a poor quality of life and a higher number of days absent from work, compared with others.12–15 Hence, successful strategies to prevent and treat headache would confer substantial benefits to afflicted individuals, as well as to society in general.

Available data support a direct association between blood pressure and the occurrence of headache.16–19 Therefore, it is reasonable to speculate that dietary factors that lower blood pressure (eg, reduced sodium intake and the DASH diet20 ,21) might also reduce the occurrence of headache. However, evidence on the relationship of headaches with sodium intake and other dietary factors is sparse, with most attention focusing on the potential role of monosodium glutamate intake.22–24 In the primary results paper of the DASH-Sodium trial, which focused on the blood pressure effects of the dietary interventions, the authors briefly comment on the occurrence of headaches in the broad context of side effects. They reported that the side effect of headache occurred in 47% of participants during the high, compared with 39 percent during the low, sodium feeding period.21 In this paper, we expand on these preliminary observations.

Methods

A detailed description of the rationale, design and methods of the DASH-Sodium trial has been published.25 Briefly, DASH-Sodium was a multicentre, randomised clinical trial, conducted between September 1997 and November 1999, designed to compare the effects on blood pressure of three levels of dietary sodium and two diet patterns. The study design incorporated a parallel, two-group, comparison of diet (DASH diet vs control diet) together with a three-period crossover of the three levels of dietary sodium intake, with a primary outcome of mean systolic blood pressure (figure 1). The three sodium levels were (1) ‘high’ (150 mmol, at 2100 kcal caloric intake), reflecting average consumption in the USA, (2) ‘intermediate’ (100 mmol) reflecting the upper limit of current recommendations for adults26 and (3) ‘low’ (50 mmol). The DASH diet is rich in fruits, vegetables and low-fat dairy products; high in dietary fibre, potassium, calcium and magnesium; moderately high in protein; and low in saturated fat, cholesterol and total fat. The control diet is typical of what many in the Western world eat.

Figure 1

DASH -Sodium trial flow diagram (BMI, body mass index; CV, cardiovascular; DM, diabetes mellitus; OTC, over the counter).

Study participants were 412 adults (age ≥22 years) with systolic blood pressure between 120 and 159 mm Hg and diastolic blood pressure between 80 and 95 mm Hg (ie, prehypertension or stage 1 hypertension). Major exclusion criteria were diabetes mellitus, evidence of active malignancy, history of cardiovascular event (angina, myocardial infarction, angioplasty or stroke), renal insufficiency (serum creatinine >1.2 mg/dL for females or 1.5 mg/dL for males), anaemia (haematocrit at least 5% below normal range), pregnancy, inflammatory bowel disease, body mass index (BMI) >40 kg/m², use of antihypertensive drugs and corticosteroids, and consumption of more than 14 alcoholic beverages per week.

Three screening visits (each separated by at least 7 days) were conducted to assess general eligibility and to collect baseline data. Following the screening visits, eligible participants started a 2-week run-in feeding period during which they ate the control diet at the high sodium level. The run-in feeding period was designed to exclude participants who were unlikely to comply with the dietary requirements and to estimate caloric requirements needed to maintain weight. Participants were then randomly assigned (generated using desktop PC at each coordinating centre) to one of the two diets using a parallel-group design, and ate each of three sodium levels (feeding periods) for 30 days each, in a randomised crossover design. Participants were not notified of their assigned dietary pattern or sodium sequence.

During feeding periods (run-in and intervention), participants were required to eat at least one meal per day on site at the clinical centre, 5 days per week, and to take food home for other meals. Participants were expected to eat all of their food and were instructed to record the type and amount of any uneaten study food. Caffeinated beverages and alcohol were limited and monitored. Individual energy intake (calorie content) was adjusted, so that each participant’s weight during each feeding period remained stable.

Data collection staff were masked to randomised sodium and diet sequence. Measurements were obtained during screening and at the end of each feeding period. Blood pressure was measured in a seated position, using the right arm of participants. Twenty-four hour urine (for analysis of sodium, potassium, urea nitrogen and creatinine) and body weight were also collected. Compliance with the feeding protocol was assessed by urinary excretion of sodium, potassium, phosphorus, urea nitrogen and creatinine, estimated from 24 h urine collections.

Symptoms (side effects), including headache, bloating, dry mouth, excessive thirst, fatigue or low energy, lightheadedness, nausea and change in taste, were collected via self-administered questionnaires (see online supplement) completed during the last 7 days of each sodium feeding period. For each symptom, potential responses were (1) ‘none’ for not experiencing any symptom, (2) ‘mild’ if symptom occurred but did not interfere with usual activities, (3) ‘moderate’ if symptom occurred and somewhat interfered with usual activities and (4) ‘severe’ if participants were unable to perform usual activities due to the symptom.

This analysis of the DASH-Sodium trial included 390 (95%) of the 412 randomised participants. Excluded participants were those with missing information on headaches in any of the three feeding periods. For the primary analysis in this study, headache was defined as ‘any headache’ (mild, moderate or severe) during the last 7 days of each feeding period. Subsequently, we report frequency of headache by severity.

The means and proportions between groups were explored using t tests and χ² tests, respectively. A non-parametric test (extension of Wilcoxon rank-sum test) was used for trends in the frequency of headache by sodium intake. Since multiple observations were obtained on each participant, we used generalised estimating equation (GEE) models,27 with a logit link and binomial error and an exchangeable covariance structure, to model the odds of a headache. The adjusted covariates used in this analysis were measured at baseline. Models were adjusted for age, sex, race, clinical site, systolic blood pressure, BMI and smoking status. The potential for carryover effects was unavoidable in this trial; however, since the experimental agent was one’s diet and participants must eat something during these intervals, statistical GEE models were also adjusted for carryover effects from the previous periods. To address the qualitative consistency and benefit-hazard profiles between participants, subgroup analysis by diet stratified by age, sex, race, obesity (BMI ≥30 kg/m² vs not) and hypertension (blood pressure ≥140/90 mm Hg vs not) status at baseline were also performed. Interactions between subgroups were tested by the addition of an interaction term to the main effects model.

Each participant provided written, informed consent.

A p value of ≤0.05 was considered statistically significant. All analyses were performed using Stata V.12.1 (Stata Corp LP, College Station, Texas, USA).

Results

The 390 participants included in our analyses were those with completed symptoms questionnaires—192 (94%) of the 204 participants assigned to the control diet and 198 (95%) of the 208 participants assigned to the DASH diet. Clinical and demographic characteristics of the two groups were similar (table 1).

Table 1

Baseline characteristics of participants in DASH-Sodium trial (number (percentage) or mean (SD))

Figure 2 displays the distribution of headaches by sodium level and assigned diet. The highest occurrence of headache was reported by participants on the control diet with high sodium (47%) and the lowest by participants on the DASH diet with low sodium level (36%). On both diets, the number of headaches reported was greatest for the high sodium level and least on the low sodium level.

Figure 2

Frequency of headache by diet and sodium level.

Among those assigned to the control diet, mean (SD) urinary sodium excretion was 141 (55), 106 (43) and 64 (37) mmol per 24 h during the high, intermediate and low sodium feeding periods, respectively. In the DASH diet group, mean (SD) urinary sodium levels were 144 (57), 107 (52) and 67 (46) mmol per 24 h during the high, intermediate and low sodium feeding periods, respectively. On each sodium level, mean urinary sodium excretion was similar in those assigned to the two diets (each p>0.05). Mean urinary potassium and urea nitrogen were higher in the DASH diet group, reflecting the higher vegetable, dairy and protein content of the DASH diet compared with the control diet, at each sodium level (table 2).

Table 2

Urinary excretion according to sodium level and diet (Mean (SD))

Table 3 shows differences in the odds of headache by diet and sodium level. Compared with the high sodium level, we observed a lower odds of any headache during the low sodium period both on the control diet (adjusted OR=0.69, 95% CI 0.49 to 0.99) and the DASH diet (adjusted OR 0.69, 95% CI 0.49 to 0.98). Although the relationship appeared graded (figure 2), there was no significant difference between the intermediate level of sodium and either the low or high sodium levels, on either diet. There was no significant association of diet pattern (DASH vs control) with headache on any sodium level. There was also no significant interaction between diet and sodium on the occurrence of headaches (p interaction >0.05). Compared with the control diet with high sodium, there was a reduced risk of a headache on the DASH diet with low sodium (adjusted OR=0.64, 95% CI 0.41 to 0.99, p=0.05).

Table 3

OR of headaches by diet and sodium sequence

While on control diet, the number of persons who reported a severe headache was 4 (2.1%) during high, 1 (0.5%) during intermediate and 1 (0.5%) during low sodium periods, respectively (p for trend =0.13). On DASH diet, the corresponding number of persons who reported a severe headache was 8 (4%) during high, 2 (1%) during intermediate and 3 (1.5%) during low sodium periods, respectively (p for trend=0.08). The frequency of severe headache was similar (p=0.3) by diet (DASH 8 (4%) and control 4 (2%)) during high sodium feeding period (table 4).

Table 4

Occurrence and severity of headache by sodium level and diet, n (%)

There was no evidence that the relationship between sodium levels and headache was modified by age, sex, race, baseline BMI or blood pressure (figure 3).

Figure 3

(A) Odds of headache (low vs high sodium) by subgroup, in the DASH diet. (B) Odds of headache (low vs high sodium) by subgroup, in the control diet.

Discussion

In this secondary analysis of the DASH-Sodium trial, which enrolled adults with prehypertension and stage 1 hypertension, a reduced dietary sodium intake was associated with a lower risk of headache, both on the control diet and the DASH diet. In contrast, the risk of headache was similar on the DASH and control diets.

The epidemiological literature on headaches in adults is limited.1 ,2 ,6 However, it is well recognised that, compared to normotensive individuals, individuals with hypertension have a higher frequency of headaches.16–19 ,28 Of note, Cooper et al17 reported a direct relationship of headaches with both systolic and diastolic blood pressure. As regards trials, in a pooled analysis that included seven double-blinded, randomised placebo controlled trials of Irbesartan therapy, Hansson et al29 found a direct relationship of diastolic blood pressure with incident headaches in 2673 patients with mild-to-moderate hypertension.

The association between dietary sodium intake and blood pressure is also well recognised.30 ,31 The DASH diet alone and in combination with reduced sodium intake lowers blood pressure in patients with or without hypertension.20 ,21 It is noteworthy that there was no significant relationship between diet pattern and headache. This suggests that a process that is independent of a blood pressure may mediate the relationship between sodium and headaches.

Our results contrast with the popular belief that a diet rich in fruits, vegetables and potassium and low in saturated and total fat may ease the frequency, or even prevent, headache.32 Several dietary factors, including fasting, alcoholic drinks, chocolate, coffee and cheese, appear to trigger vascular headache (cluster or migraine) in adults.33–36 In some studies, an increased intake of monosodium glutamate is associated with the occurrence of headaches.22–24 However, a recent review concluded that evidence on the relationship of sodium glutamate intake and headaches is inconsistent.37 In one study of 200 adults (mean age 37.7 years, 81% females), monosodium glutamate was identified as a trigger for migraine headache in only 5 (2.5%) of study participants.36 However, data on the relationship between sodium intake and any form of headaches are sparse.

The results of this study provide encouraging evidence in support of dietary recommendations to lower sodium intake: recommendations which are currently based on the relationship of sodium intake with blood pressure. The daily intake of sodium in adults living in the USA is already in excess of their physiological need and for many individuals, is much higher than the highest level tested in this study.38 ,39 Our results also support the recent WHO guidelines for reducing sodium intake to less than 87 mmol/day40 and American Heart Association guidelines for reducing sodium intake to 65 mmol/day.31

Strengths of our study include its randomised controlled design comparing two diets using a parallel design and a three-period crossover of three levels of dietary sodium (high, intermediate and low). Dietary intake during the feeding periods was closely monitored and vigorous efforts were made to promote adherence with assigned diets. The participants of this study were healthy, non-institutionalised, racially diverse, middle-aged and older-aged men and women. Hence we believe that these results are applicable to a large fraction of adults.

Our study also has limitations. Information on the prevalence of headache at baseline from eligible participants was lacking. In addition, there was no information about the type of headache (tension, cluster or migraine) experienced by study participants. However, we suspect that most of the headaches were tension headaches. Whether a reduced sodium intake can prevent vascular headache is unknown. Second, the instrument was administered just once in each feeding period and does not allow calculation of an event rate, such as person-days of headaches. Third, these are secondary, post hoc analyses from a trial that was not explicitly designed to test the effects of dietary factors on headaches. Nonetheless, a rigorously controlled feeding study designed to test the effects of dietary factors on occurrence of headaches would be extremely expensive and logistically challenging. Fourth, our results likely underestimate the relationship of sodium intake with headaches. The range of sodium intake was relatively narrow—the highest sodium group in our trial actually corresponds to the average in the USA and is much lower than the intake in many countries, particularly in Asia. Self-report of symptoms is inherently imprecise and could bias results to the null, given that a validated instrument was not used for patient-reported headache.

In conclusion, a reduced sodium intake was associated with significantly lower risk of headache, while diet patterns had no effect on the risk of headaches. A reduced dietary sodium intake offers a novel approach to prevent headache in adults. Additional studies are needed to replicate these findings and to explore mechanisms that mediate the association between sodium intake and headache.

Page not found |

Page not found |

404. Page not found

Monthly archive

MonTueWedThuFtSaSun

13141516171819

20212223242526

2728293031

12

12

1

3031

12

15161718192021

25262728293031

123

45678910

12

17181920212223

31

2728293031

1

1234

567891011

12

891011121314

11121314151617

28293031

1234

12

12345

6789101112

567891011

12131415161718

19202122232425

3456789

17181920212223

24252627282930

12345

13141516171819

20212223242526

2728293031

15161718192021

22232425262728

2930

Archives

Tags

Settings
for visually impaired

90,000 Hyponatremia is… What is Hyponatremia?

Hyponatremia is a condition in which the concentration of sodium ions in the blood plasma falls below 135 mmol / L (normally 150 mmol / L). Hyponatremia is caused by a wide range of pathologies.

Symptoms

Most patients are asymptomatic, but symptoms of an illness causing hyponatremia may appear.

Severe hyponatremia can cause osmotic redistribution of water from blood plasma to cells in the body, including brain cells.Typical symptoms in this case include dizziness, vomiting, headaches, and general malaise. As hyponatremia deepens, confusion, numbness (stupor), and coma may occur. Since dizziness itself is a stimulus for the excretion of ADH (antidiuretic hormone), there is the possibility of a self-reinforcing cycle of effect (positive feedback).

Reasons

An abnormally low sodium concentration in the blood should be investigated in relation to the osmolarity of the blood plasma and the volume of the patient’s extracellular fluid.

In most cases, hyponatremia is associated with a decrease in plasma osmolarity. The vast majority of cases of hyponatremia in adults are associated with increased activity of ADH (antidiuretic hormone) (ADH is a hormone that regulates the balance of water in the body, but not salt). Therefore, a patient with hyponatremia can be considered as a patient with increased ADH activity. The task of the doctor in this case is to determine the cause of the increased activity of ADH.

In patients with a loss of fluid (blood) volume, the secretion of ADH is increased because the reduction in blood volume is a natural stimulus for the secretion of ADH.As a result, the patient’s kidneys retain water and excrete highly concentrated urine. Treatment is simple – restoring the patient’s blood volume, thus turning off the ADH excretion signal.

In some patients with hyponatremia, the blood volume is normal. In these patients, the increased level of ADH activity and subsequent water retention may be due to physiological causes of ADH retention such as pain or dizziness. Another possible cause is inappropriate ADH syndrome (SIADH).In this syndrome, ADH is released continuously, at levels significantly higher than normal and is most often a side effect of certain drugs, lung problems (such as pneumonia or abscess), brain diseases, or certain types of cancer (most often small cell lung carcinoma).

The third group of patients with hyponatremia is characterized by the presence of peripheral edema (edema). The fluid in the edematous tissue does not participate in circulation and stagnates. As a result, the free blood volume decreases, and this in turn leads to the excretion of ADH.Treatment of such patients should be aimed at eliminating the causes of edema. In many cases, this is not easy at all, since the real causes can be cirrhosis of the liver or heart disease, the treatment of which cannot be considered simple.

Patients using diuretics should be considered separately. These medicines increase the excretion of fluid in the urine and, as a result, decrease in blood volume. As mentioned above, a decrease in blood volume is a stimulus for ADH and renal water retention.

The current surge in mortality from hyponatremia is associated with excess water consumption under the influence of methylenedioxymethamphetamine (“ecstasy”). Also Almond et al. ^[1] found mild hyponatremia in 13% of the Boston marathon runners, and life-threatening (blood sodium levels below 120 mmol / L) in 0.6%. Runners who put on weight during a run due to excess water intake are at greatest risk.

Hypoosmolar hyponatremia

When the osmolarity of blood plasma is low, the volume of extracellular fluid in the body can be in one of three states:

Low volume

The loss of water is accompanied by the loss of sodium.

It is necessary to treat the cause of the disease and carry out intravenous injections of saline solution. It is important to note that the sudden restoration of blood volume causes the cessation of ADH excretion. Accordingly, a normal urine output begins in the kidneys. This can lead to a sudden and significant increase in serum sodium concentration and increases the risk of myelinosis of the central pons. This syndrome is characterized by severe damage to the nervous system, often irreversible.

Due to the risk of myelinosis, patients with low fluid volumes may require injections of not only physiological (saline) solution, but also water.These measures ensure a smoother rise in serum sodium concentration as blood volume increases and a smoother decrease in ADH levels.

Normal volume

• Syndrome of inappropriate secretion of ADH (SIADH)
• In some cases of psychogenic polydipsia (insatiable thirst)

In the case of SIADH, demeclocycline may be beneficial as an ADH antagonist.

Increased volume

At the same time, water is retained in the body.

Helps to limit the patient’s water intake.

Severe hyponatremia may result from several hours of intense physical activity in high ambient temperatures, such as running or walking in the desert, or endurance exercise without adequate electrolyte intake by the athlete.

Notes

1. ↑ Almond CS et al. (2005) Hyponatremia among runners in the Boston Marathon. N Engl J Med, 352 (15): 1550-6.PMID 15829535

See also

References

90,000 Hyponatremia associated with the use of drugs of various groups.

Dear colleagues! The Center for Drug Safety Assessment (CEBS) reminds you that the use of certain drugs may be associated with a risk of developing hyponatremia with corresponding clinical consequences.

Dear colleagues!

The Center for Evaluation of the Safety of Medicines (CEBS) reminds you that with the use of certain drugs there may be a risk of developing hyponatremia with the corresponding clinical consequences.

Thus, the Australian regulatory authority (TGA) analyzed reports of the development of drug-related hyponatremia (MP), which showed that this complication of drug therapy was most often caused by the following groups of drugs:

• diuretics,
• antidepressants,
• antiepileptic drugs,
• antihypertensive drugs.

Hyponatremia is a pathological condition characterized by a decrease in the concentration of sodium ions in the blood plasma below 135 mmol / L.Most patients are asymptomatic, but severe hyponatremia can lead to a redistribution of water from the blood plasma to cells in the body, including brain cells. Typical symptoms in this case are dizziness, vomiting, headaches and general malaise. As hyponatremia worsens, psychiatric symptoms, numbness (stupor), and coma may occur.

During the period from January 2009 to 2011, the TGA received 136 reports of the development of hyponatremia in patients.

Table 1 provides information about drugs that caused the development of electrolyte disturbances.

Analysis of 136 messages showed that hyponatremia, in more than half of the cases, developed in patients over 70 years old – 76 cases (56%).

The most common clinical manifestations of hyponatremia were confusion, dizziness, dehydration, nausea and vomiting, but in some cases hyponatremia was asymptomatic and laboratory data were used as a diagnostic criterion.

It should also be borne in mind that mental disorders can also be associated with hyponatremia. So, a case of their development was revealed in a 71-year-old patient while taking carbamazepine and indapamide. The content of sodium ions in the blood plasma of this patient decreased to 114 mmol / l.After the cancellation of both drugs, the reverse development of CPD was noted.

These data indicate the need to take into account the likelihood of developing the complications described above when prescribing drugs that can cause hyponatremia.

.