Dietary Sodium Intake and Incidence of Congestive Heart Failure in Overweight US Men and Women:  First National Health and Nutrition Examination Survey Epidemiologic Follow-up Study FREE

DESPITE THE dramatic decline in mortality from coronary heart disease (CHD) and stroke in the US population, mortality from congestive heart failure (CHF) has been increasing for several decades.^1- 2 Approximately 4.7 million Americans have CHF and 47 000 die of CHF in the United States each year.¹ The identification of modifiable risk factors for CHF, which provided a basis for intervention strategies, has become an important challenge for public health and clinical medicine.³

Several cross-sectional epidemiologic studies^4- 7 have suggested a significant and independent association between dietary sodium intake and left ventricular hypertrophy. In the Treatment of Mild Hypertension Study,⁴ dietary sodium intake was estimated from 2 consecutive overnight urine collections and left ventricular structure was assessed by M-mode echocardiography in 511 men and 33 women with mild hypertension. After adjustment for age, sex, race, body mass index, previous medication, alcohol consumption, cigarette smoking, physical activity, and systolic blood pressure, a difference of 25 mmol per 8 hours in urinary sodium excretion was associated with a 41% (95% confidence interval, 22%-63%) increase in the odds of left ventricular hypertrophy. Messerli and colleagues⁵ reviewed 9 cross-sectional studies in which the association between dietary sodium intake and left ventricular mass was assessed. The correlation coefficients for the relationship varied from 0.22 to 0.61, with P<.05 in each of the studies. Animal studies^8- 9 also indicate that high sodium intake is an independent risk factor for the development of cardiac hypertrophy. However, there are sparse data on the relationship between dietary sodium intake and the risk of CHF from prospective cohort studies.

We took advantage of the large sample size and prolonged follow-up experience of participants in the first National Health and Nutrition Examination Survey (NHANES I) Epidemiologic Follow-up Study (NHEFS) to examine the relationship between dietary sodium intake and risk of CHF in a nationally representative sample of the US general population.

In NHANES I, a multistage, stratified, probability sampling design was used to select a representative sample of the US civilian noninstitutionalized population aged 1 to 74 years. Details of the study design, sampling methods, response rate, and data collection have been previously published.^10- 11 Certain population subgroups, including those with a low income, women of childbearing age (25-44 years), and persons 65 years or older were oversampled.

The NHEFS is a prospective cohort study of NHANES I participants who were aged 25 to 74 years when the survey was conducted between 1971 and 1975.^12- 15 Of the 14 407 NHEFS participants who were in this age range at their baseline examination, we excluded 3059 who lacked 24-hour dietary recall information by study design; 2 who lacked sodium intake information; 176 who had a history of CHF at their baseline examination, defined as ever having been told by a physician that they have had heart failure or having used any medication for a "weak heart" during the 6 months before their baseline interview; and 395 who were consuming a low-salt diet at baseline because of heart disease or hypertension. Among the remaining participants, 413 (4%) were lost to follow-up. After these exclusions, experience from 5233 nonoverweight and 5129 overweight participants was available for analysis. Overweight was defined as a body mass index (calculated as weight in kilograms divided by the square of height in meters) of 25.0 or greater for men and women.

Baseline data collection included a medical history, a standardized medical examination, a dietary history, laboratory tests, and anthropometric measurements.^10- 11 A single 24-hour dietary recall was conducted by trained NHANES I personnel using a standardized protocol and 3-dimensional food portion models. Information on frequency, but not amount, of salt added in the kitchen or at the table was collected. The dietary recall questionnaires were later coded by interviewers using nutrient information from the US Department of Agriculture Handbook No. 8 or other resources. Dietary sodium and caloric intake were calculated for each participant by the National Center for Health Statistics, Hyattsville, Md. Frozen serum samples were sent to the Centers for Disease Control and Prevention, Atlanta, Ga, for measurement of serum total cholesterol levels. Blood pressure measurements, body weight, and height were obtained using standard protocols. The baseline questionnaire on medical history included questions about selected health conditions and medications used for these conditions during the preceding 6 months. Data on educational level, physical activity, and alcohol consumption were obtained by interviewer-administered questionnaires. Baseline information on smoking status was obtained in a random subsample of 6913 participants who underwent a more detailed evaluation at the time of their examination.^10- 11 For the remaining study participants, information on smoking status at baseline was derived from responses to questions regarding lifetime smoking history, which were administered at follow-up interviews conducted from 1982 to 1984 or later.^16- 17 The validity of information on smoking status obtained using this approach has been documented.^16- 17 Baseline diabetes was defined as ever having been told by a physician that the participant had this condition. Hypertension was defined as a systolic blood pressure of 160 mm Hg or higher and/or a diastolic blood pressure of 95 mm Hg or higher and/or use of antihypertensive medications. Coronary heart disease was defined as ever having been told by a physician that the participant experienced a heart attack or having a diagnosis with an International Classification of Diseases, Ninth Revision (ICD-9), code between 410 and 414. Valvular heart disease was defined as having a diagnosis with an ICD-9 code between 394 and 397 or of 424 at the baseline medical examination.

Follow-up data were collected between 1982 and 1984 and in 1986, 1987, and 1992.^12- 15 Each follow-up examination included tracking a participant or his or her proxy to a current address; performing an in-depth interview; obtaining hospital and nursing home records, including pathology reports and electrocardiograms; and, for decedents, acquiring a death certificate. Incident cardiovascular disease was based on documentation of an event that met prespecified study criteria and occurred between the participant's baseline examination and last follow-up interview. Mortality from cardiovascular disease was based on death certificate reports. The validity of study outcome data from both sources has been documented.¹⁸

Incident CHF was based on 1 or more hospital and/or nursing home stays in which the participant had a discharge diagnosis with an ICD-9 code between 428.0 and 428.9 or a death certificate report in which the underlying cause of death was recorded using an ICD-9 code between 428.0 and 428.9. Incident CHD was based on 1 or more hospital and/or nursing home stays in which the participant had a discharge diagnosis with an ICD-9 code between 410 and 414 or a death certificate report in which the underlying cause of death was an ICD-9 code between 410 and 414. The date of record for incident events was identified as the date of the participant's first hospital admission with an established study event or the date of death from a study event in the absence of hospital or nursing home documentation of such an event.

Because a statistically significant interaction between dietary sodium intake and overweight status on CHF was detected, all analyses were stratified by overweight status. The quartiles of dietary sodium intake were calculated using the total sample (nonoverweight and overweight participants). For each baseline characteristic, the mean value or corresponding percentage of study participants was calculated by overweight status. The cumulative incidence of CHF by quartile of sodium intake was calculated using the Kaplan-Meier method, and differences in cumulative rates were examined by the log-rank test for trend.^19- 20 Cox proportional hazards models were used to explore the relationship between dietary sodium intake and cardiovascular disease risk.²¹ History of CHD was modeled as a time-dependent variable in Cox proportional hazards models. Age was used as the time scale for all time-to-event analyses.²² Cox proportional hazards models were stratified by birth cohort using 10-year intervals to control for calendar period and cohort effects.²² Sodium intake was assessed as a categorical (quartile) and a continuous variable. Methods to estimate variances that take into account sample clustering and stratification of the NHANES I sample were used in Cox proportional hazards models.²² Sex differences in the relationships between sodium-calorie ratio and cardiovascular diseases and all-cause mortality were tested using interaction terms in Cox proportional hazards models. Because there were no significant differences, men and women were pooled in the main analysis.

Table 1 shows the baseline characteristics of the study participants by weight status. Compared with the nonoverweight study participants, those who were overweight were older; were more likely to be men, African American, and have a high school education; had a higher mean systolic blood pressure and serum total cholesterol level; had a higher prevalence of hypertension, diabetes mellitus, and CHD; were less likely to engage in recreational physical activity; and had a lower prevalence of current cigarette smoking and alcohol consumption. In addition, dietary intake of sodium, potassium, calcium, and total calories (energy) was lower in the overweight compared with the nonoverweight participants.

During 85 035 person-years of follow-up from 1971 through 1992, 413 CHF events were documented in the nonoverweight participants. The cumulative incidence of CHF was not significantly associated with baseline sodium intake in the nonoverweight participants. For example, the cumulative incidence of CHF at age 90 years, adjusted for total calorie intake, was 35.6%, 42.9%, 37.4%, and 47.4% among patients within the first, second, third, and fourth quartiles of dietary sodium intake, respectively (P = .48 for trend). During a total of 80 265 person-years of follow-up in the overweight participants, 679 CHF events were documented. The cumulative incidence of CHF at age 90 years, adjusted for total calorie intake, was 44.0%, 43.6%, 53.6%, and 63.1% among overweight participants within the first, second, third, and fourth quartiles of dietary sodium intake, respectively (P = .02 for trend).

Relative risks and 95% confidence intervals for CHF according to baseline quartiles of dietary sodium intake in the nonoverweight and overweight participants are presented in Table 2. Dietary sodium intake was not significantly associated with risk of CHF in the nonoverweight study participants. In contrast, dietary sodium intake was significantly associated with risk of CHF among overweight individuals. In age-, race-, sex-, and total calorie intake–adjusted analyses, CHF incidence was positively and significantly associated with the corresponding trend in dietary sodium intake in the 5129 participants who were classified as overweight. After additional adjustment for educational level, physical activity, cigarette smoking, alcohol consumption, systolic blood pressure, serum total cholesterol level, history of diabetes mellitus and valvular heart disease, and time-dependent history of CHD, the association remained largely unchanged and statistically significant. Likewise, additional adjustment for dietary intake of potassium and calcium did not materially change the results. Similar findings were obtained when quartile of sodium-calorie ratio was used as the independent variable and when history of hypertension was used as an adjustment variable instead of systolic blood pressure (data not shown).

Table 3 presents multivariate relative risks of CHF associated with a 100-mmol/d or per 1743 kcal increase in sodium intake by overweight status, where sodium intake (100 mmol/d) and sodium-calorie ratio (100 mmol per 1743 kcal, average energy intake in the study population) were treated as continuous variables. There were borderline statistically significant interactions between sodium intake and overweight status on CHF. Dietary sodium intake was significantly associated with an increased risk of CHF incidence in overweight, but not in nonoverweight, persons. Furthermore, the relative risks were similar when either sodium-calorie ratio or absolute sodium intake was used as the independent variable.

To the best of our knowledge, the present study is the first prospective investigation to document a strong and independent relationship between dietary intake of sodium and increased risk of CHF among overweight persons. Previous information on dietary sodium intake and CHF has come from cross-sectional epidemiologic studies and animal experiments.^4- 9 Among those studies, a positive association between dietary sodium intake and measures of left ventricular structure was a consistent observation.

The findings from the present study have important public health and clinical implications. The incidence of and mortality from CHF have been increasing relentlessly in the US population during recent decades.^1- 2 With an increasingly older population and progressive improvements in survival following acute myocardial infarction, it is almost inevitable that CHF will continue to be an important public health challenge in the foreseeable future. Because of the high mortality associated with CHF, it is important to identify modifiable risk factors and develop effective strategies for the prevention of CHF in the general population. We previously reported that cigarette smoking, overweight status, physical inactivity, and a history of hypertension, diabetes mellitus, and CHD were important modifiable risk factors for CHF in the study population.³ The present findings suggest that a high dietary intake of sodium is an independent risk factor for CHF and that a reduced sodium intake may play an important role in the prevention of CHF in the community.

A moderately low intake of dietary sodium (<2400 mg/d) has been recommended for all Americans by several organizations and government agencies.²³ The beneficial effect of a moderately low intake of dietary sodium includes decreasing the risk of hypertension and cardiovascular disease.²⁴ Observational epidemiologic studies^25- 26 have documented that a high intake of dietary sodium is associated with an elevated blood pressure and an increased risk of hypertension. Randomized controlled trials^27- 28 have also demonstrated that a reduced dietary intake of sodium lowers blood pressure in hypertensive and normotensive persons. Several prospective cohort studies^29- 30 have indicated that a high dietary intake of sodium increases the risk of stroke and CHD. Our study findings provide additional information to support recommendation for a moderately low intake of dietary sodium in the US population.

Our study indicates that the increased risk of CHF associated with high dietary sodium intake presented in overweight persons only. It has been suggested that overweight persons may be more sodium sensitive because of enhancing renal tubular sodium reabsorption and sodium retention.^31- 32 In a study of 60 obese and 18 nonobese adolescents, Rocchini et al³³ found that blood pressure was more sensitive to dietary sodium intake in overweight than nonoverweight adolescents and that this increased sodium sensitivity was reduced after weight loss. Increased sodium sensitivity in overweight persons may explain why we identified an independent positive relationship between dietary sodium intake and risk of CHF. Several previous epidemiologic studies^29- 30 also indicated that the association between dietary sodium intake and risk of cardiovascular disease was much stronger in overweight compared with nonoverweight persons. Overweight is a common and important risk factor for CHF.³ According to data from the third NHANES, conducted from 1988 to 1991, 54.5% of US adults 20 years or older were overweight, as defined by a body mass index of 25 or higher.³¹ Particularly troublesome, the prevalence of obesity has increased progressively during recent decades in all race and sex groups in the United States.³⁴ To reduce the CHF risk in overweight patients, weight loss and sodium reduction should be recommended. For persons with difficulty losing body weight, greater attention to reductions in sodium intake may be appropriate.

An excess dietary intake of sodium might induce sodium and water retention, especially in persons who are more sensitive to sodium intake, such as obese patients.^31- 32 Sodium and water retention plays an important role in the pathogenesis of hypertension and CHF.^35- 36 A high dietary intake of sodium might originate CHF by increasing blood pressure (pressure overload) or extracellular fluid (volume overload).

There are several potential limitations of the present study. The dietary sodium intake was estimated by a single 24-hour dietary recall. This may result in misclassification of usual sodium intake at the individual level. In addition, the dietary recall method used in NHANES I may have underestimated sodium intake because it did not include quantitative data on discretionary use of salt in cooking or at the table settings, which account for about 15% to 30% of the sodium intake in Western societies.^37- 38 Likewise, NHEFS participants might have underreported their dietary total caloric intake, which further underestimated sodium intake.³⁹ In addition, repeated measurements of dietary sodium intake during the follow-up period are not available from the NHEFS. These measurement errors would tend to bias our relative risk estimates toward 1 (effect to 0) in univariate models. Besides a possibly lower sensitivity to sodium intake in nonoverweight compared with overweight persons, our finding of a null association in nonoverweight persons may have been attributed to the measurement errors in sodium intake. Another limitation of our study was the fact that study participants were followed up in a passive rather than an active fashion. Because CHF status was identified by information obtained from hospital discharge diagnoses or death certificates, the cumulative incidence of CHF may be underestimated. However, there is no reason to believe that CHF outcome measurements would differ by baseline dietary intake of sodium. Because of the nature of the passive follow-up, echocardiographic or electrocardiographic data were not available and left ventricular dysfunction could not be studied.

Our study has several important strengths. The findings can be generalized to overweight persons in the US general population because the NHEFS cohort is a random sample of this population. In addition, temporal relationships can be established with confidence because dietary sodium intake was measured at baseline and subsequent incidence of CHF was assessed over an average duration of 19 years. The fact that follow-up experience was available for more than 96% of study participants further enhances the validity of our findings.

In conclusion, our study results suggest that a high dietary intake of sodium is an independent risk factor for CHF. Furthermore, these findings suggest that sodium reduction may play an important role in the prevention of CHF in the US general population. Future prospective cohort studies with better measurements of usual dietary sodium intake (ie, multiple measurements of 24-hour urinary excretion of sodium at baseline and during follow-up visits) and CHF outcome are needed to confirm our findings.

Accepted for publication November 20, 2001.

This study was supported by grant R03 HL61954, and in part by grant R01 HL60300, from the National Heart, Lung, and Blood Institute, Bethesda, Md. The NHANES I NHEFS has been developed and funded by the National Center for Health Statistics, Hyattsville, Md; the National Institute on Aging, Bethesda; the National Cancer Institute, Bethesda; the National Institute of Child Health and Human Development, Bethesda; the National Heart, Lung, and Blood Institute, Bethesda; the National Institute of Mental Health, Rockville, Md; the National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda; the National Institute of Arthritis and Musculoskeletal and Skin Diseases, Bethesda; the National Institute of Allergy and Infectious Diseases, Bethesda; the National Institute of Neurological and Communicative Disorders and Stroke, Bethesda; the Centers for Disease Control and Prevention, Atlanta, Ga; and the US Department of Agriculture, Washington, DC.

Corresponding author and reprints: Jiang He, MD, PhD, Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, 1430 Tulane Ave, Mail Box SL18, New Orleans, LA 70112 (e-mail: jhe@tulane.edu).