Salt sensitivity is defined as a change in blood pressure in response to changes in salt and water homeostasis. Found in 73% of hypertensive and 36% of normotensive blacks, it is generally considered a hallmark of hypertension in blacks. The higher prevalence of salt sensitivity in blacks compared with whites suggests a genetic influence on this trait, but there is little direct evidence of heritability. We determined the extent to which salt sensitivity is correlated in black families and estimated the heritability of this phenotype. Black families were recruited through a hypertensive proband. Both hypertensive and normotensive adults were phenotyped with respect to salt sensitivity with an intravenous sodium-loading, furosemide volume-depletion protocol. Salt sensitivity was defined as the difference between sodium-loaded and volume-depleted blood pressure. We enrolled 20 families, comprising 30 parent-offspring pairs and 115 adult sibling pairs. Age-adjusted familial correlations ranged from .33 to .44, .19 to .37, and .12 to .21 for mean arterial and systolic and diastolic pressure responses to the salt sensitivity maneuver, respectively. Corresponding heritability estimates were 0.26 to 0.84, 0.26 to 0.74, and 0.004 to 0.24, respectively. These data strongly suggest a heritable component of salt sensitivity.
Salt sensitivity is defined as a change in blood pressure (BP) in response to changes in salt and water homeostasis. Found in 73% of hypertensive and 36% of normotensive blacks, it is generally considered a hallmark of hypertension in blacks.1 Understanding the mechanism or mechanisms underlying salt sensitivity may greatly enhance our understanding of the pathophysiology of essential hypertension in black men and women. There is an obvious need to investigate potential mechanisms of essential hypertension in the black population. Compared with white Americans, black Americans experience a higher prevalence of hypertension, suffer more severe consequences, and appear to differ in their target-organ response to hypertension.2 The higher prevalence of salt sensitivity in blacks than in whites suggests a genetic influence on this trait, but there is little direct evidence of heritability. For instance, there is significant familial correlation of the BP response to sodium restriction.3 In addition, renal sodium handling,4 5 the BP response to volume depletion,4 and the BP response to a volume-expansion/depletion protocol identical to the one used in the present study4 are more closely correlated in monozygotic than dizygotic twins. However, it is unclear to what extent these traits represent salt sensitivity in its traditional meaning. Furthermore, it is unclear to what extent these observations in predominantly white populations apply to salt sensitivity in blacks. In this study, we determined the extent to which salt sensitivity is correlated in black families and we estimated its heritability. Our data suggest a genetic contribution to salt sensitivity.
Ascertainment and Eligibility
This study was approved by the Institutional Review Board of Duke University Medical Center, and all subjects provided written informed consent. Black families were ascertained through a single hypertensive proband. Using a sequential sampling scheme, we first enrolled all available adult first-degree relatives (ie, siblings, offspring, and parents) of each proband and then enrolled all available adult second-degree relatives (grandparents, aunts, uncles, and half-siblings) of the proband.
All probands were hypertensive by design. Other participants were hypertensive or normotensive. Hypertension status was determined by averaging duplicate seated BP measurements taken on each of two clinic visits at least 1 week apart. Hypertension status was established at least 2 weeks after all BP-lowering medication had been discontinued and before measurement of salt sensitivity. All BPs were measured by trained personnel using a mercury sphygmomanometer and an appropriately sized cuff.6 Subjects were considered hypertensive if mean diastolic BP was greater than 95 mm Hg. Subjects were considered normotensive if there was no history of hypertension and if mean systolic BP was less than 140 mm Hg and mean diastolic BP was less than 90 mm Hg in the absence of any BP-lowering medications. We were concerned that isolated systolic hypertension represents a pathophysiological entity distinct from other forms of essential hypertension. Therefore, to increase the clinical homogeneity of our study population, we excluded individuals with elevated systolic but normal diastolic BP.
Subjects were excluded if they were younger than 18 years. In addition, for safety purposes, subjects were excluded if there was a history of malignant or accelerated hypertension, if diastolic BP was greater than 120 mm Hg, if there was a contraindication to discontinuing BP medications, or if there was a history of furosemide allergy. Subjects were also excluded if there was evidence of impaired renal function (ie, serum creatinine >133 μmol/L [1.5 mg/dL], urinary protein excretion >500 mg/d, or active urinary sediment).
Other exclusion criteria were myocardial infarction, cerebrovascular accident, or transient ischemic attack within the previous 6 months; congestive heart failure by history, physical examination, or chest radiograph; volume depletion; severe peripheral vascular disease; or inability to give informed consent or to comply with the study protocol.
To confirm that reported parent-offspring relationships were biological, we typed parents and offspring with a battery of red blood cell markers and antigens designed to detect at least 70% of mistaken parentage.7 8 We excluded from analysis any subjects in whom parentage could not be confirmed.
Definition of Salt Sensitivity
Salt sensitivity was measured by the intravenous sodium-loading/furosemide volume-depletion protocol of Grim et al.9 Subjects were evaluated on the Duke Clinical Research Unit. This protocol was approved by the Duke University Medical Center Institutional Review Board for Clinical Investigations, and all subjects gave signed informed consent. At the time of admission to the Clinical Research Unit, subjects had unrestricted sodium intake. Dietary intake was estimated from a 24-hour urine collection obtained on the day before admission to the Clinical Research Unit.
Sodium Chloride Loading
After baseline BP measurements, normal saline (0.9%) was infused via the brachial vein at a rate of 2 L over 4 hours (8 am to noon). The total sodium intake during this 24-hour period was approximately 500 mmol (intravenous plus dietary). Dietary potassium intake was 70 mmol/d.
Starting at 8 am on the day after sodium loading, dietary sodium intake was limited to 10 mmol. Dietary potassium remained at 70 mmol, and subjects were allowed no more than 25 mL/kg water. Anthropometric measures (body weight, height, skinfold thickness) were used for estimation of lean body weight. Furosemide was then administered at a dose of 1 mg/kg lean body weight, divided into three oral doses over the day. Each subject was weighed before the second furosemide dose. If weight loss due to diuresis exceeded 3% of baseline (total) body weight, the third furosemide dose was reduced to avoid excessive volume depletion.
Using the criteria established by Weinberger et al,1 we used two BP measurements to define salt sensitivity: a saline-loaded BP measured at the end of the saline infusion and a volume-depleted BP taken 24 hours after the first furosemide dose. Each BP measurement consisted of the average of three seated measurements taken by trained personnel6 with a mercury sphygmomanometer using the same arm and appropriately sized cuff. Subjects were seated for at least 5 minutes before the first BP recording; subsequent recordings were taken at 30-second intervals. In addition, for safety purposes, BP was measured every hour throughout the saline-loading and volume-depletion procedures.
Subjects ended the salt-loading procedure before completion if systolic BP exceeded 180 mm Hg, if diastolic BP exceeded 120 mm Hg, or if the subject developed symptoms related to elevated BP. Subjects ended the volume-depletion procedure before completion if they developed symptomatic hypotension that could not be resolved by decreasing the furosemide dose. If a subject ended the procedure early because of hypotension or hypertension, the last BP before termination was used in the salt-sensitivity calculations.
Salt sensitivity was defined as the difference between sodium-loaded and volume-depleted pressures. Therefore, for most individuals, the measure of salt sensitivity represented a decrease in systolic and diastolic BPs and mean arterial pressure (MAP). Although the decrease in pressure was expressed relative to (ie, as a percentage of) the sodium-loaded pressure, similar results were obtained when absolute values were used. These continuous variables were used in correlation analyses. MAP was calculated from the formula [(2×Diastolic BP)+Systolic BP]/3.
Familial correlations were determined with the FCOR program of the Statistical Analysis of Genetic Epidemiology program (SAGE, 1994) for parent-offspring (interclass) and sibling (intraclass) correlations. There were insufficient data for evaluation of spouse-spouse pairs. Analyses are reported in two ways: calculated with equal weight attributed to each pedigree and with equal weight attributed to each sibpair. In addition, because of a positive association between salt sensitivity with age (data not shown), we performed these analyses with and without age adjustment.
We used these familial correlations to estimate the heritability of salt sensitivity. In this context, heritability is the proportion of the total variance in a quantitative trait that is due to additive genetic factors.10 Typically, a heritability estimate of zero suggests little or no genetic effect, whereas an estimate of 1 suggests a large genetic effect. (There are exceptions to this rule: the heritability will be zero if a trait is inherited but there is no genetic variation in the sample.) Intermediate heritability estimates reflect varying contributions of both environmental and genetic factors. In this study, heritability is estimated as twice the parent-offspring correlation.11
We enrolled subjects from 20 black families, including 30 parent-offspring pairs and 115 adult sibling pairs. The mean age was 36 years, 66% of participants were women, and 27% had hypertension. Overall, average (±SD) systolic and diastolic BPs and MAP were 118±15, 79±12, and 92±12 mm Hg, respectively. Mean Quetelet index was 31±9 kg/m2; urinary excretions of sodium and potassium were 142±60 and 41±55 mmol/d, respectively. Plasma renin activity was 2.2±1.7 ng/mL per hour.
Fifteen individuals experienced side effects to the salt-sensitivity procedure, but only 1 withdrew early. This subject developed severe hypertension during the sodium-loading procedure. Five subjects experienced symptomatic hypotension during the volume-depletion procedure. In each case, the total furosemide dose was decreased (see “Methods”), and the subject was able to complete the study. Two subjects developed nausea after furosemide treatment, and 4 subjects experienced nonspecific complaints.
As noted above, our definition of salt sensitivity was the relative change in BP. Table 1⇓ indicates that among all subjects, systolic pressure decreased by an average of 8.6±8% (range, 44% decrease to 11% increase), diastolic pressure by 0.9±9% (20% decrease to 24% increase), and MAP by 4.0±9% (24% decrease to 18% increase) from the sodium-loaded state (ie, at the end of the saline infusion) to the volume-depleted state (ie, 24 hours after the first furosemide dose).
Correlation coefficients were calculated with an age adjustment because of the known relationship between salt sensitivity and age,12 which was also observed in our population (data not shown). Age-adjusted familial correlations are presented in Table 2⇓. Parent-offspring correlations calculated with equal weight to pedigrees were clearly nonzero for MAP and systolic BP responses to sodium loading and depletion (.42 and .37, respectively). When age-adjusted correlations were recalculated with equal weight to pairs as opposed to pedigrees, the r values were generally lower for parent-offspring correlations and largely unchanged for sib-sib correlations. “Equal weight to pairs” means that each pair is given equal weight in the calculation of the sum of squares and sum of products in the Pearson product-moment correlation. Thus, a family with four siblings will contribute six times more information to the sibling correlation and twice as much information to the parent-offspring correlation than a family with two siblings. In contrast, “equal weight to pedigrees” means that each family is given equal weight in the calculation of the correlations regardless of the size of the family. The fact that we observed higher correlations with equal weight to pedigrees than to pairs (reflected in increased estimates of the heritability, below) may reflect underlying genetic heterogeneity of salt sensitivity. For instance, if larger families happen to have lower familial correlations than smaller families, then the estimates of heritability will be lower for equal weight to pairs than for equal weight to pedigrees. This may be due to just a few large families with lower correlations. However, the approximate threefold increase in correlations for MAP response and systolic BP response suggests that the heritability for these traits is, at least in some families, quite high.
In this setting, it is unclear whether age adjustment will give the most accurate information about heritability. Adjusting for age removes the variability in salt sensitivity that is due to aging itself, but the unadjusted values allow for the possibility that the genetic effect actually increases with age (ie, delayed and variable age of onset of the genetic trait). Nonetheless, unadjusted correlations are also shown in Table 2⇑ and are about the same or somewhat higher than the age-adjusted correlations.
We doubled the age-adjusted parent-offspring correlations to provide estimates of the heritability of these traits. Table 3⇓ shows that depending on whether equal weight was given to pairs or to pedigrees, 26% to 84% of the observed variability in the MAP response to sodium loading and volume depletion could be attributed to genetic factors. These estimates were similar for systolic salt sensitivity (26% to 74%) but considerably higher than the range of estimates for heritability of diastolic salt sensitivity (0.4% to 24%).
Many investigators hypothesize that salt sensitivity reflects an underlying predisposition to hypertension. This hypothesis is based on a very large body of suggestive evidence, beginning with cross-cultural comparisons demonstrating increasing prevalence of hypertension with increasing salt intake.13 Yet even in populations with high salt intake, 60% of the population remains normotensive, suggesting that some individuals are susceptible to the BP effects of habitual salt intake.14 Salt sensitivity is generally considered a hallmark of hypertension in blacks and is present in 73% of hypertensive blacks.1 This racial predilection has led to speculation that salt sensitivity is inherited. Our data suggest that the majority of the variation in the salt sensitivity of MAP and systolic BP may be due to additive genetic effects.
Lower estimates of heritability were obtained for the diastolic BP response. This is somewhat surprising because estimates of the heritability of BP itself (ie, without manipulation of volume status) are very similar for systolic BP (0.60 to 0.70) and diastolic BP (0.52 to 0.61).15 16 However, there is some evidence that systolic and diastolic BPs are under independent genetic influences.17 Alternatively, some proportion of these subjects could have clinically undetected diastolic dysfunction secondary to hypertension, which might alter the diastolic response to sodium manipulations and mask genetic influences on this trait.
Before this study, there was little direct evidence of the heritability of the salt-sensitivity phenotype. A genetic basis for this phenotype has been suggested by the association of salt sensitivity with haptoglobin phenotype.18 More direct evidence comes from a study by Whitfield and Martin5 in which renal sodium handling (measured as the fractional excretion of sodium) was more highly correlated in monozygotic than dizygotic twins, leading to a heritability estimate of 0.5. However, their study did not include measurement of the effect of changes in sodium intake on BP and therefore cannot be directly applied to the study of salt sensitivity. The strongest prior evidence that salt sensitivity is inherited comes from two twin studies evaluating the BP response to low sodium diet3 and to furosemide volume depletion.4 In the former study, the age-adjusted correlation between MAP response in dizygotic pairs was .36 and in monozygotic pairs was .68. The higher correlation in monozygotic twins (who are essentially identical genetically) than in dizygotic twins (who have on average 50% genetic homology) strongly supports genetic influences on this phenotype. In this same study, parent-offspring correlations led to estimates of heritability ranging from 0.34 to 0.74 (ie, 34% to 74% of the variability in salt sensitivity was attributed to genetic influences). In the latter study, with methods identical to the present study, heritability of both systolic and diastolic BP responses to furosemide volume depletion was greater than 0.7 (the authors do not report correlations for the response to sodium loading).
The authors of both of these studies defined salt sensitivity as the reduction in BP with sodium depletion—an alternative definition of salt sensitivity that may or may not represent the same phenomenon as the definition we used. Several methods have been published for measuring and defining salt sensitivity. We chose the intravenous saline-loading/furosemide volume-depletion protocol originally described by Grim et al in 19799 because it is performed under controlled circumstances (ie, on a Clinical Research Unit) and because there is a large body of data in which salt sensitivity is measured by this method.19 Nonetheless, these two definitions of salt sensitivity are correlated in our own population (r=.54, P=.0001), and the estimates of heritability are similar to those in the study of Miller et al,3 suggesting that our data in families and the twin data of Miller et al represent consistent evidence of the heritability of BP response to changes in sodium homeostasis.
Both of the twin studies described above are limited to white subjects. Salt sensitivity is found in 29% to 56% of white hypertensive individuals and may play the same role in this population as it does in blacks.9 20 However, heritability estimates are often population specific.21 Therefore, any future studies of the genetic basis of salt sensitivity in black families should be based on heritability estimates in a black population.
Despite the high rates of salt sensitivity in blacks with hypertension, salt sensitivity is not a surrogate for hypertension in this population, nor is it likely to be the sole pathophysiological factor in the development of hypertension in any population. However, a large body of evidence suggests that salt sensitivity leads to hypertension through vascular, hormonal, and/or cation transport mechanisms. For instance, salt sensitivity is associated with increased forearm vascular resistance, which increases further with sodium loading (no such change is observed in sodium-resistant individuals).22 23 Thus, salt sensitivity may be associated with increased peripheral vascular resistance, the hallmark of essential hypertension. Salt sensitivity is also associated with blunted renin and aldosterone responses to sodium depletion.24 Sodium loading in salt-sensitive individuals does not reduce angiotensin II–induced vasoconstriction,25 nor does it cause the normal suppression of plasma norepinephrine levels.26 Taken together, these hormonal findings suggest that sodium sensitivity may be related to abnormal regulation of the angiotensin axis and/or the adrenergic nervous system. Salt sensitivity is also correlated with sodium-lithium countertransport.27 The level of activity of this membrane transport system is genetically determined and may affect peripheral resistance through effects on the cytosolic calcium content of vascular smooth muscle cells.28 Thus, several lines of evidence suggest a pathophysiological connection between salt sensitivity and hypertension.
Even if we speculate that the relationship between salt sensitivity and high BP is epiphenomenal and not causal, further characterization of the salt-sensitivity phenotype may lead to a better understanding of BP regulation, with obvious implications for our understanding of essential hypertension. In addition, the salt-sensitivity phenotype may define a subpopulation of hypertensive individuals that is genetically more homogeneous than the general population of hypertensive blacks. In the study of multifactorial genetic traits, subdividing populations into more homogeneous groups can markedly facilitate the search for disease genes.29 Our data lay the necessary foundation for a comprehensive search for hypertensive genes in the black population through the investigation of the salt-sensitivity phenotype.
This research was supported by Grant-in-Aid 92012180 from the American Heart Association and by grant RO1-HL-50176 from the National Institutes of Health (NIH). Additional support was provided by M01-RR-30 National Center for Research Resources, General Clinical Research Centers Program, NIH.
Reprint requests to Dr Laura P. Svetkey, Duke Hypertension Center, 3024 Pickett Rd, Durham, NC 27705. E-mail email@example.com.
- Received March 5, 1996.
- Revision received April 2, 1996.
- Revision received May 27, 1996.
Weinberger MH, Miller JZ, Luft FC, Grim CE, Fineberg NS. Definitions and characteristics of sodium sensitivity and blood pressure resistance. Hypertension. 1986;8(suppl II):II-127-II-134.
Miller JZ, Weinberger MH, Christian JC, Daugherty SA. Familial resemblance in the blood pressure response to sodium restriction. Am J Epidemiol. 1987;126:822-830.
Grim CE, Miller JZ, Luft FC, Christian JC, Weinberger MH. Genetic influences on renin, aldosterone, and the renal excretion of sodium and potassium following volume expansion and contraction in normal man. Hypertension. 1979;1:583-590.
Whitfield JB, Martin NG. Genetic and environmental causes of variation in renal tubular handling of sodium and potassium: a twin study. Gen Epidemiol. 1985;2:17-27.
Widmann F, ed. Technical Manual of the American Association of Blood Banks. 9th ed. Arlington, Va: American Association of Blood Banks; 1985: chaps 8-10.
Silver H, ed. Paternity Testing. Arlington, Va: American Association of Blood Banks; 1978: chap 3.
Grim CE, Luft FC, Fineberg NS, Weinberger MH. Responses to volume expansion and contraction in categorized hypertensive and normotensive man. Hypertension. 1979;1:476-485.
Vogel F, Motulsky AG. Human Genetics: Problems and Approaches. 2nd ed. New York, NY: Springer-Verlag; 1986:181-184.
Weinberger MH, Fineberg NS. Sodium and volume sensitivity of blood pressure: age and pressure change over time. Hypertension. 1991;18:67-71.
Feinleib M, Garrison RJ, Fabsitz R, Christian JC, Hrubec Z, Borhani NO, Kannel WB, Rosenman R, Schwartz JT, Wagner JO. The NHLBI twin study of cardiovascular disease risk factors: methodology and summary of results. Am J Epidemiol. 1977;106:284-295.
Avolio A. Genetic and environmental factors in the function and structure of the arterial wall. Hypertension. 1995;26:34-37.
Weinberger MH, Miller JZ, Fineberg NS, Luft FC, Grim CE, Christian JC. Association of haptoglobin with sodium sensitivity and resistance of blood pressure. Hypertension. 1987;10:443-446.
Sullivan JM. Salt sensitivity: definition, conception, methodology, and long-term issues. Hypertension. 1991;17(suppl I):I-61-I-68.
Falkner B, Kushner H. Effect of chronic sodium loading on cardiovascular response in young blacks and whites. Hypertension. 1990;15:36-43.
Khoury MJ, Beaty TH, Cohen BH. Genetic approaches to familial aggregation. In: Fundamentals of Genetic Epidemiology. New York, NY: Oxford University Press; 1993:200-377.
Mark AL, Lawton WJ, Abboud FM, Fitz AE, Connor WE, Heistad DD. Effects of high and low sodium intake on arterial pressure and forearm vascular resistance in borderline hypertension: a preliminary report. Circ Res. 1975;36(suppl I):I-195-I-198.
Koolen MI, Bussemaker-Verduyn den boer E, van Brummelen P. Clinical, biochemical, and haemodynamic correlates of sodium sensitivity in essential hypertension. J Hypertens. 1983;1(suppl 2):21-23.
Weder AB. Membrane sodium transport and salt sensitivity of blood pressure. Hypertension. 1991;17(suppl I):I-74-I-80.
Blaustein MP. Sodium ions, calcium ions, blood pressure regulation, and hypertension: a reassessment and a hypothesis. Am J Physiol. 1977;232:C165-C173.
Lander ES, Schork NJ. Genetic dissection of complex traits. Science. 1994;265:2037-2048.