Sodium Sensitivity of Blood Pressure: A New Prognostic Factor in Hypertension

We recently showed the sodium sensitivity of blood pressure (BP) to be an important cardiovascular prognostic factor in hypertension [1]. Theoretically, there is a close relationship between sodium sensitivity and glomerular capillary hypertension [2, 3], and indeed the glomerular capillary hydraulic pressure is proved to be elevated in all animal models examined in which the BP is sensitive to changes in sodium intake [2]. Furthermore, in all animal models with a sodium-sensitive type of hypertension, progressive glomerulosclerosis occurs, resulting in renal failure [2]. Even in this type of essential hypertension, the glomerular pressure is estimated to be elevated [4, 5]. Microalbuminuria, a marker of glomerular capillary hypertension [6, 7], is also known to be augmented in this type of hypertension [6, 8]. In addition, we also showed that the BP failed to fall during night in patients with the sodium-sensitive type of essential hypertension, and they manifested as nondippers who had a specific circadian BP rhythm without nocturnal dip [9, 10]. It is also known that in essential hypertension with nondipper type of circadian BP rhythm, the urinary albumin excretion rate is augmented [11]. Recently, it has been reported that microalbuminuria [12, 13]and ambulatory BP pattern [14]both can be independent predictors of the prognosis in hypertension. Insulin resistance [15, 16, 17, 18, 19, 20], dyslipidemia [19], and sympathetic nerve activity, reported to be accelerated in sodium-sensitive states, together with two factors stated above, may participate with a high incidence of cardiovascular complications in the sodium-sensitive type of hypertension. Here, I would like to summarize the new entity that sodium sensitivity of the BP itself is an important cardiovascular prognostic factor.Dahl [21]developed by selective inbreeding two strains of rats: one, a salt-sensitive ‘S’ strain, exhibited hypertension with salt feeding, and the other, a salt-resistant ‘R’ strain, remained normotensive despite a high salt intake. Furthermore, replacing the kidneys of ‘R’ rats with the kidneys from ‘S’ rats led to hyertension when these ‘R’ rats were fed a high-salt diet, while transplantation of kidneys of ‘R’ into ‘S’ rats prevented the development of hypertension [22]. These experiments afforded strong evidence that salt can indeed induce hypertension and that a genetic defect of hypertension in ‘S’ rats is expressed in the kidneys. This defect, thought to be involved in influencing renal sodium excretory capability, was soon confirmed by Tobian et al. [23]who showed a less steep pressure-natriuresis relationship in perfused kidneys isolated from Dahl ‘S’ rats.Bartter’s group [24, 25]showed for the first time that even in human essential hypertension, the BP response to a high salt intake was of two types, not unlike that in Dahl rats. Since then, sodium sensitivity of BP has been noted especially in the aspects where mechanisms causing hypertension have been different from others.Bartter’s group [24, 25]gave a low-sodium diet containing 9 mmol of sodium/day for 7 days and then switched to a high-sodium diet containing 249 mmol/day for 7 days. Although the change in mean arterial pressures (MAP) induced by the sodium load seemed to show a regular and continuous distribution from small to large without a clearcut point of separation, these authors arbitrarily divided the response into two types: patients were considered ‘sodium sensitive’ when the MAP was elevated more than 10%, while lesser responders were regarded as ‘not sodium sensitive’. Some authors followed Bartter’s protocol and arbitrary definitions, whereas others modified amount [5, 26, 27, 28]and duration [5, 27, 28]of each sodium intake period and also took 5% [28]or 10 mm Hg [5, 27, 29]instead of 10% elevation in blood pressure as evidence for a ‘sodium-sensitive’ response. The criteria for the definition of sodium sensitivity have thus varied markedly [30].These inconsistencies made it clear that ‘salt sensitivity’ should be considered two-dimensionally based on BP elevation as well as magnitude of sodium load [31, 32, 33, 34, 35, 36], since the BP elevation can be greater when the sodium intake is greater. Also, it became clear that the magnitude of sodium load should be evaluated in each patient as the difference in urinary sodium excretion (UNaV) in the steady state of sodium balance at each level of sodium intake.An approximately linear relationship has been recognized [37, 38, 39, 40]between UNaV and systemic MAP (fig. 1). This pressure-natriuresis relationship, which is believed to be determined by the kidneys, is called renal function curve by Guyton [37, 38, 39]. When the amount of dietary sodium intake is plotted on the UNaV axis of the pressure-natriuresis curve, the BP value of the crossing-point of the above two lines is the equilibrium pressure [37, 38, 39], indicating that UNaV becomes equal to the amount of sodium intake only at this equilibrium BP.Although it is ideal to generate data over a larger continuum to assess this curve more precisely, the arterial pressure-natriuresis relationship has ususally been found to be very near to linear within the experimental range of sodium intake used in animals [37, 41, 42]and within the wide range of sodium intake in normal humans [43, 44]as well as in hypertensive subjects [45, 46, 47]. Assuming a linear relationship, the UNaV can now be expressed as a first-order function of MAP using the extrapolated x-intercept, A, and the slope, B, as [2, 48, 49, 50, 51, 52, 53]:UNaV = B × (MAP – A)(1)Only when the BP reached the level of equilibrium MAP, MAPeq, the UNaV must be equal to the amount of sodium intake, QNa:B × (MAPeq – A) = QNa(2)Solving equation 2, the equilibrium BP can be expressed as:Here, C is the proportional constant to the amount of sodium intake, defined as 1/B. The equations 3 and 4 illustrate several important aspects of hypertension. First, the steady state BP can be expressed as the sum of two components: the non-sodium-sensitive component determined by the x-intercept of the pressure-natriuresis curve and the sodium-sensitive one determined by the product of the reciprocal of the slope multiplying the amount of sodium intake. Therefore, now it is clear that the reciprocal of the slope reflects the sodium sensitivity of the BP, since C (= 1/B) is the proportional constant converting the amount of sodium intake into BP elevation dependent on sodium intake [53]. In addition, it is also clear that the steady state BP, which is independent of cardiac output and total peripheral resistance, will be elevated only by either an increase in the x-intercept (causing non-sodium-sensitive hypertension) or a decrease in the slope (causing sodium-sensitive hypertension).Once the sodium sensitivity of blood pressure is defined as discussed above as the reciprocal of the slope of the pressure-natriuresis relationship, it can be obtained as the ratio of the change in MAP (mm Hg) over the change in UNaV (mmol/day), as shown in figure 1 and equation 5 [32, 53], by placing patients on two different levels of sodium intake.Since BP and UNaV rate are both measured in the steady state after achievement of sodium balance at each level of sodium intake, this index can readily be compared for results obtained at different levels of sodium intake. In some studies, ‘sodium sensitivity’ has been examined by observing the BP fall in response to reducing sodium intake [33, 34, 54]. Since, as discussed above [45, 46, 47], the relationship between UNaV and systemic MAP is approximately linear, the new definition of the sodium sensitivity index provides insights into true sodium sensitivity irrespective of whether the sodium intake is increased (sodium loaded) or reduced (sodium depleted) [41, 45, 46, 47, 55], independently of the magnitude of the change in the amount of sodium intake. In addition, this definition can be used widely to assess the relationship between sodium sensitivity and other parameters or phenomena [1, 2, 4, 9, 10, 15, 34, 47]without separating patients into two groups: non-sodium-sensitive and sodium-sensitive patients. I believe that equation 3 for the first time clarifies the sodium sensitivity of BP theoretically, because the reciprocal of the slope of pressure-natriuresis curve converts the amount of sodium intake into BP elevation dependent on sodium intake.From the foregoing, it is evident that the steady state BP as well as its sodium sensitivity are determined by the pressure-natriuresis relationship. Therefore, if specific renal determinants of the pressure-natriuresis relationship are known, then renal mechanisms of sodium sensitivity are clarified necessarily. The x-intercept, i.e., the degree of the shift along the BP axis, is determined by the sum of the pressure drop from heart to glomeruli plus the opposing pressures at glomeruli [51, 52, 56]. The main factor seems to be the preglomerular vascular resistance from heart to glomeruli, mainly due to the afferent arteriolar resistance. The slope, the reciprocal of the sodium sensitivity, is determined by the difference between glomerular (mainly due to the whole-kidney ultrafiltration coefficient) and tubular (rate of sodium reabsorption) factors [52].Let me now attempt to classify various forms of hypertension based on renal mechanisms (table 1) [57, 58, 59, 60, 61, 62]. Theoretically, the pressure-natriuresis relationship can be affected in two different ways to cause hypertension [37, 38, 39, 40, 49, 50, 52]: either a parallel shift toward a higher BP level along the BP axis or a depression in the slope (see figure 1). As already discussed, the parallel shift of the pressure-natriuresis relationship induces non-sodium-sensitive hypertension, while the depressed slope induces sodium-sensitive hypertension. Since the x-intercept of the pressure-natriuresis relationship is determined by the preglomerular vascular resistance from heart to glomeruli [51], the parallel shift should be due to the increased preglomerular vascular resistance. Examples for this mechanism are Goldblatt hypertension in animal models [63, 64]and renovascular hypertension in humans [33, 65, 66]. It is well known that these forms of hypertension are based on an increase in vascular resistance due to renal artery stenosis. There is ample evidence to suggest that spontaneously hypertensive rats of Kyoto strain and human essential hypertension result from increased afferent arteriolar resistance [37, 51, 56, 65, 67, 68].On the other hand, as already seen, the depressed slope of the pressure-natriuresis relationship induces sodium-sensitive hypertension. The slope is determined by the difference between the whole-kidney ultrafiltration coefficient (KF) and tubular sodium reabsorption. Therefore, a depressed slope results from a decrease in KF, an increase in tubule sodium reabsorption, or both. KF is the product of the single-glomerulus ultrafiltration coefficient (Kf) and the number of glomeruli [69, 70]. In Dahl salt-sensitive rats and during the early phase of glomerulonephritis, the reduced Kf per glomerulus may contribute to a decreased value of KF [57, 58], although there are also data suggesting a reduced nephron number [59]and an increased tubular sodium reabsorption [60]in Dahl rats. Examples of decreased KF due to the reduced number of glomeruli are Milan hypertensive rats [61], oligomeganephronia [71, 72, 73], congenital solitary kidney [74, 75, 76, 77, 78], and extensive renal ablation in humans in general [79, 80, 81, 82]and hypertension in black subjects [62]. It should be stressed again that the slope of the pressure-natriuresis relationship is determined by glomerulotubular balance of sodium, instead by the KF itself. When the number of glomeruli and nephrons is reduced, both KF and rate of tubular sodium reabsorption will be reduced, with litte affecting this balance. Thus, the reduced number of nephrons seems to have less impact on the sodium sensitivity than the reduced Kf per glomerulus [37, 40].An enhanced tubular sodium reabsorption clearly contributes to the genesis of hypertension in desoxycorticosterone acetate salt hypertension and primary aldosteronism [33]. Likewise, during the early phases of diabetes mellitus, increases in tubular sodium reabsorption due to enhanced Na-glucose cotransport may be the principal mechanism for hypertension [83, 84, 85].In sodium-sensitive states, whether the KF is reduced or the tubular sodium reabsorption is enhanced, the glomerular capillary hydraulic pressure rises to compensate for the impairments in sodium excretory capability. In other words, it may be considered that both sodium sensitivity and glomerular hypertension reflect adaptations necessary to overcome a defect in the capacity of the kidneys to excrete sodium [86]. Elevation in glomerular capillary pressure leads to a syndrome of glomerular hyperfiltration/hypertension characterized by progressive proteinuria, systemic hypertension, eventual glomerular sclerosis, and eventual renal failure [87, 88, 89]. If the linkage between sodium sensitivity of BP and glomerular hypertension is as proposed, renal failure may be expected to be more common in sodium-sensitive types of hypertension than in non-sodium-sensitive types. Indeed, animal models characterized by increased sodium sensitivity of BP are proved to show glomerular hypertension and more profound renal injury [5, 58, 59, 62, 90, 91, 92, 93, 94, 95, 96].As discussed above, high sodium sensitivity of BP implicates the presence of glomerular capillary hypertension. Microalbuminuria, which is a marker of glomerular capillary hypertension [6, 7], has now been considered a prognostic indicator not only for renal dysfunction, but also for cardiovascular diseases [12, 13, 97, 98, 99, 100]. Furthermore, in high-sodium-sensitive states, including essential hypertension [9, 10, 101, 102, 103], primary aldosteronism [104], chronic renal failure [105, 106, 107, 108, 109], and diabetes mellitus [110], the nocturnal BP fails to fall, resulting in nondippers. Recently, it has been reported that nondipping, associated with serious target organ damage such as left ventricular hypertrophy, microalbuminuria, and cerebrovascular disease [11, 111, 112, 113, 114, 115], can be an independent predictor of prognosis in essential hypertension [14]. Furthermore, a linkage between high sodium sensitivity and insulin resistance is suggested [15, 16, 17, 18, 19, 20], the latter of which has been considered as one of the important atherosclerotic cardiovascular risk factors [116, 117, 118, 119, 120]. Therefore, in the sodium-sensitive state of hypertension, several risk factors or markers seem clustered together, leading to future cardiovascular morbid events [1, 15, 19, 20, 121, 122].We, therefore, examined whether sodium sensitivity of BP might be a new independent prognostic factor in essential hypertension. One hundred and fifty-six patients with essential hypertension were placed on high (12–15 g NaCl/day) and low (1–3 g/day) sodium diets for 1 week each to determine the sodium sensitivity and were then followed up retrospectively for 7.3 ± 4.3 years, until the end points such as ischemic heart disease and stroke occurred [1]. Patients whose MAP was lowered more than 10% by sodium restriction were considered sodium-sensitive patients (n = 62), while the remaining were considered non-sodium-sensitive patients (n = 94). Between both groups, there were no differences in age, diabetes and sex distribution, body mass index, serum cholesterol and creatinine concentrations, or MAP. Left ventricular hypertrophy was recognized more frequently in the sodium-sensitive group and in current smokers in the non-sodium-sensitive group. During the follow-up period, cardiovascular events occurred in 14 cases (including 3 fatal events) of the non-sodium-sensitive group and in 17 cases (5 fatal) of the sodium-sensitive group. The rates of cardiovascular morbid events per 100 patients years were 2.0 in the non-sodium-sensitive group and 4.3 in the sodium-sensitive group (fig. 2). The cardiovascular event-free survival curve was significantly worse in the sodium-sensitive group. The Cox proportional-hazards model identified sodium sensitivity, MAP at discharge, and smoking as risk factors for cardiovascular morbidity, while contributions from other classical risk factors were not significant. These results show that the cardiovascular morbidity was higher in the sodium-sensitive type than in the non-sodium-sensitive type of essential hypertension. Thus, it was concluded that sodium sensitivity appeared to be a new and strong cardiovascular risk factor in essential hypertension, independent of classical factors.In the above study, 31 cardiovascular events were documented during the follow-up period, among which 20 events were stroke [1]. On the other hand, it has been reported that hypertensive subjects with a high plasma renin activity, especially white men, are more likely to have subsequent myocardial infarction [123, 124]. Since, in general, a high plasma renin activity is associated with a low sodium sensitivity and a low plasma renin activity with a high sodium sensitivity, the above reports seem to contradict our hypothesis. However, the authors [123, 124]clearly stated [123, 124]that stroke was not associated with a high plasma renin activity and speculated there to be differences in etiology of the two conditions, coronary heart disease and stroke, or alternately a different relation of the renin-angiotensin system to coronary and cerebral vasculatures. Therefore, the difference between results of previous reports and our hypothesis may be ascribed to differences in cardiovascular events whether coronary heart disease or stroke or differences in race whether white or Japanese. Our results, described above, may support their speculation on the differences in the two conditions. It is also interesting to note here that sodium sensitivity is more common in people of Japanese origin, and strokes are more frequently seen than ischemic heart disease. Similarly, people of African origin are more sodium sensitive than Caucasians [125, 126], and strokes are known to be more common in the former [127, 128]. Thus, there may be a linkage between sodium sensitivity and stroke among Japanese and African-Americans.Primary aldosteronism is known as one of the most typical forms of sodium-sensitive hypertension [33, 104, 129]. Recently, we found 9 cases (15.5%) of stroke in 58 patients with primary aldosteronism [97], being consistent with other reports [130], which may support our proposal for an association between sodium-sensitive type of hypertension and stroke.Sodium restriction has been considered one of the most important lifestyle modifications as well as a therapeutic strategy in the field of hypertension and renal diseases. Nevertheless, there is no consensus on its importance in arresting the cardiovascular complications. As discussed above, sodium restriction in patients with high sodium sensitivity may lower systemic and glomerular capillary pressures, reduce proteinuria, and normalize circadian BP rhythm for a nondipper to a dipper pattern. Therefore, sodium restriction may relieve these risks and prevent cardiovascular diseases. I believe sodium sensitivity of BP must be noted as an important prognostic predictor for cardiovascular diseases. Probably, sodium sensitivity reflects loss of renal functional reserve and severity of hypertension that can not be predicted by classical factors such as blood pressure itself, risk factors, and target organ damage.This work was supported by Research Grants for Cardiovascular Diseases (A-1997-9 and C-1999-4) from the Ministry of Health and Welfare of Japan as well as by grants from Salt Science Research Foundation, Takeda Science Foundation, and Japan Cardiovascular Research Foundation.