|Home | About | Journals | Submit | Contact Us | Français|
We studied the characteristics of insulin resistance in 19 normotensive and 25 hypertensive subjects who underwent an acute protocol for determination of salt-sensitivity of blood pressure. Hypertensive subjects were older and more obese, with higher creatinine, lipids and aldosterone than normotensive volunteers. They also had higher glucose and insulin levels with a marked decrease in insulin sensitivity (HOMA2-S index). Once all participants were classified into salt-sensitive (SS) and salt-resistant (SR) groups, most of these differences were no longer present. In contrast, SS had classical characteristics of this phenotype (higher percentage of blacks, suppressed plasma renin, increased aldosterone-to-renin ratio and blunted renin and aldosterone responses to changes in salt balance). Despite similar insulin levels, HOMA2-S was significantly lower in SS than SR. Salt loading did not change HOMA2-S in SS or SR. In contrast, salt depletion, by significantly increasing glucose and insulin of SR, decreased their HOMA2-S to the levels observed in SS. Correlates of insulin resistance in SR included age, triglycerides, BMI, mean arterial pressure, aldosterone and epinephrine. However, only BMI and aldosterone remained as significant predictors in multivariate analyses. Correlates of insulin resistance in SS were mean arterial pressure, epinephrine and norepinephrine, all remaining as significant predictors in multivariate modeling. Our data confirm that salt sensitivity of blood pressure is associated with insulin resistance, suggest that salt restriction may be beneficial in SS but perhaps detrimental in SR subjects, and uncover possible differences in mechanisms of insulin resistance between SS and SR, with implications for pharmacological therapy.
Salt sensitivity of blood pressure (SSBP) is a phenotype defined by changes of systemic arterial pressure that parallel changes in sodium balance. In contrast, salt-resistant (SR) animals or humans excrete a sodium load without changes in blood pressure (BP); i.e., without resorting to pressure natriuresis. Although many abnormalities in natriuretic and antinatriuretic systems have been described in SSBP, there is no definitive answer regarding its actual etiology. It is however widely accepted that the phenotype is genetically determined, because inbreeding experiments have created homogenous strains of salt-sensitive (SS) rodents. In humans, SSBP is normally (not dichotomously) distributed1 and SS subjects have increased cardiovascular morbidity and mortality.2
One quarter to one third of normotensive humans are SS whereas up to 80 percent of hypertensive subjects are, depending on age and race. The reason by which SSBP carries prognostic implications independent of BP is not known but it has been suggested that future development of hypertension in normotensive subjects and diminished nocturnal dipping of arterial pressure3,4 may play a role. Also, insulin resistance is a feature present in SS rodents5,6 and humans7–10. Although all essential hypertension is an insulin resistant state (high prevalence of carbohydrate metabolism abnormalities in hypertensive subjects11), there is evidence that insulin resistance of SS subjects is either more prevalent or more severe than in SR, consistent with observations in rodents, and this might account for the prognostic impact of SSBP.
Vasoactive systems involved in hypertension (renin angiotensin system12,13, aldosterone14, or catecholamines15) may produce insulin resistance via molecular pathways that inhibit insulin signaling or via hemodynamic actions that impair glucose uptake in peripheral tissues, particularly skeletal muscle. There is little information on whether insulin resistance of SS subjects differs from that of SR in its characteristics or underlying mechanisms. This is particularly important because there is also controversy on the effects of varying salt intake on insulin sensitivity in humans15–23. We therefore decided to study correlates of insulin resistance in normotensive volunteers and essential hypertensive patients, analyzing data separately after classifying them into SS and SR groups. Our studies suggest that the predictors of insulin resistance are different in SS vs SR. This may have implications in terms of current guidelines regarding salt intake and also in terms of selection of pharmacological therapy.
Normotensive volunteers and essential hypertensive patients were recruited into an IRB-approved inpatient protocol to study effects of acute salt loading and salt depletion on BP. All of them provided informed consent. Out of a total of 53 subjects, 5 were excluded for lack of insulin data, and 6 for having diabetes; results for the remaining 42 are reported here.
We have previously described the protocol in detail 24. Briefly, after obtaining demographic information and routine laboratory data, hypertensive subjects discontinued therapy for two weeks. All subjects maintained their usual salt intake until admission to the hospital for an overnight stay. Baseline BPs and heart rates were obtained by automated ambulatory monitors (Spacelabs 90207) from 7 to 8 am the following morning, followed by salt loading on the first day of study (160 mEq NaCl diet plus 2L normal saline infused from 8 AM to 12 PM) and salt depletion on the second (10 mEq NaCl diet plus three 40-mg doses of oral furosemide). BPs and heart rates were recorded throughout the study with the ambulatory monitors, every 15 minutes from 6 AM to 10 PM and every 30 minutes overnight. Average systolic BPs from noon (end of saline on the first day or second dose of furosemide on the second day) to 10 PM (bedtime) were used for classification of subjects into SS or SR groups; a fall of 10 mm Hg in systolic BP from salt loading to salt depletion was the cutoff. Body weights were recorded at baseline and daily at 7 am, before interventions. BMI was calculated as weight in Kg divided by height in m2.
Blood samples for routine tests, plasma renin activity (PRA), aldosterone, insulin and catecholamines (epinephrine and norepinephrine) were obtained at baseline (before saline infusion at 8 AM of the first day) and on the mornings after the days of salt loading and salt depletion. PRA, aldosterone and insulin were measured by radioimmunoassay and catecholamines by radioenzymatic assay.
Insulin sensitivity (S) and beta-cell function (B) were assessed using the HOMA2 model with the paired fasting glucose and insulin levels of each day employing the HOMA2 Calculator v2.2 of the Diabetes Trials Unit, University of Oxford (www.dtu.ox.ac.uk).25
Data are presented as mean ± standard errors of the mean. Skewed data were log transformed. Comparisons between groups were made with unpaired t tests. Changes in variables between periods of the protocol were analyzed with paired t tests. Differences between proportions were evaluated by chi-square testing. Correlation coefficients were calculated with Pearson’s method, pooling the data for the three days of the protocol in each patient. These tests, single linear regressions and stepwise mixed multivariate analyses for model fitting were performed with JMP software (version 3.0.2, SAS Institute). A probability less than 5% was used to reject the null hypothesis.
The 23 hypertensive and 19 normotensive subjects were 42.6±1.4 years old, 81% female and 38% black. Obesity (BMI > 30 Kg/m2) was present in 64% of subjects.
The left columns of Table 1 show that hypertensive subjects were older, with a larger percentage of black subjects. In addition to the expected difference in BP, their BMI, heart rate, serum creatinine, plasma aldosterone and epinephrine were higher than those in normotensive volunteers, whereas creatinine clearance was lower. Because PRA of hypertensive subjects was somewhat (not significantly) higher than that of normotensives, aldosterone-to-renin ratios were not different between groups. The prevalence of obesity was greater in hypertensive than normotensive subjects (78.3 vs 47.4%, chi-square <0.05, not shown). Consistent with this, glucose, total and LDL cholesterol, triglycerides and plasma insulin were significantly higher, whereas insulin sensitivity index was significantly lower in hypertensive than in normotensive subjects.
The prevalence of SSBP was somewhat higher in hypertensive (43.5%) than in normotensive (36.8%) subjects but its magnitude (assessed by the reduction in systolic BP from the salt-loading to the salt-depleted days) did not differ between SS (−13.5±1.9 vs −15.9±1.5 mmHg), or SR (−2.2±1.2 vs −3.3±1. 3) normotensive and hypertensive subjects.
The right columns of Table 1 show data after all subjects were reclassified into SS and SR groups. Several of the differences between normotensive and hypertensive subjects were no longer observed between SS and SR, since a similar percentage of hypertensive and normotensive subjects belonged to each of these groups. Hence, age, BMI, heart rate, creatinine, creatinine clearance, lipids, epinephrine, glucose and insulin were similar between SS and SR. In contrast, SS were characterized by larger percentage of black subjects, higher diastolic BP, lower PRA, higher aldosterone-to-renin ratios and increased BP variability compared with SR. Despite similar insulin levels, SS had higher (not statistically significant) blood glucose levels, leading to significantly lower insulin sensitivity index (i.e., greater insulin resistance) in SS than in SR.
Figure 1 shows effects of salt loading (comparing baseline data with those after salt infusion and high salt diet) and salt depletion (comparing data after salt loading with those after low salt diet and furosemide) on renin, aldosterone and catecholamines in SS and SR. PRA was suppressed by the salt load and stimulated by salt depletion in both groups, as expected, but the values of each day were lower in SS than in SR. Stimulation of PRA by salt depletion in SS was blunted compared to that in SR (0.6±0.2 vs 1.3±0.3 ngAI/L/sec, p<0.03). Aldosterone levels were suppressed by salt loading and stimulated by salt depletion in SR, whereas in SS, its suppression by salt was not significant and its stimulation by salt-depletion was of lesser magnitude than in SR (142±50 vs 268±48 pmol/L, p<0.05). Aldosterone-to-renin ratios (not shown) did not change significantly in response to salt loading in either group, consistent with concomitant suppression of PRA and aldosterone by salt loading. In contrast, they were significantly diminished by salt depletion in both groups, because of greater stimulation of PRA than of aldosterone by this intervention. Finally, in every phase of the experiment, aldosterone-to-renin ratios were higher in SS than SR subjects, reflecting the lower PRA of the former.
Plasma norepinephrine and epinephrine (Figure 1) did not differ between SS and SR on any day of the experiment. Salt loading reduced the levels of both, but this change was significant for epinephrine in SS only. In contrast, both catecholamines were significantly stimulated by salt-depletion, to the same extent in SS and SR.
Figure 2 shows the effects of salt loading and depletion on metabolic parameters. Glucose, insulin levels, and beta-cell function (estimated by HOMA2-B) were not significantly different between SR and SS during the baseline and salt loaded days. However, slightly higher glucose in SS during those two days (about 0.47 mmoL/L higher than those in SR) resulted in markedly diminished insulin sensitivity (HOMA2-S) compared to SR. After salt depletion, significant increases in both glucose and insulin levels took place in SR, leading to a 24% decrease in insulin sensitivity in this group, reaching a level indistinguishable from that observed in SS during all periods of the study. Salt depletion produced a modest significant increase in glucose in SS, but the slight further diminution of their HOMA2-S was not significant. In contrast to these changes in insulin sensitivity, HOMA2-B was not altered by salt depletion in either group. In other words, SS subjects were more insulin resistant, compared to SR, during baseline and after salt loading, but salt depletion induced insulin resistance in SR subjects, equalizing HOMA2-S in both groups, without modifying beta-cell function.
There was no correlation between the effects of salt depletion on systolic BP (i.e., the salt sensitivity index) and the baseline insulin sensitivity of the patient. An apparent trend (r=0.30, p=0.06) was due to SS having larger falls in systolic BP and lower HOMA2-S, whereas the opposite was true in SR. However, there was no relationship between salt sensitivity and insulin resistance when each group was analyzed separately.
Figure 3 depicts metabolic and renin-angiotensin system univariate correlates of HOMA2-S in SR (open circles) and SS (closed circles). Regression lines are plotted for the group with significant correlations only. Older age, magnitude of obesity, higher triglycerides and higher log aldosterone correlated with greater insulin resistance (i.e., lower HOMA2-S) in the relatively more insulin sensitive SR group only. None of these variables correlated with insulin sensitivity in the more insulin resistant SS subjects. Log PRA and log aldosterone-to-renin ratio did not correlate with HOMA2-S in either group (not shown).
Figure 4 depicts hemodynamic and catecholamine univariate correlates of HOMA2-S in SR and SS. Significant correlations are indicated by solid regression lines for SR and by dashed ones for SS. Higher mean arterial pressures correlated with increasing insulin resistance in both groups, consistent with the observation that hypertensive subjects were more insulin resistant than normotensive ones (Table 1). The slope in SS was flatter than that of SR (0.51±0.20% reduction of HOMA2-S per each mmHg increase in MAP in SS, versus 1.25±0.29% in SR, p<0.02). Hence, at the same BP level, if within the normal range, SS subjects were more insulin resistant than SR. In contrast, with high BP values, the magnitude of insulin resistance at any level of BP was similar in both groups.
Higher plasma epinephrine also correlated with increasing insulin resistance (lower HOMA2-S) in both groups, but the relationship was stronger (r value) in SS. The slopes were parallel (104±43% reduction of HOMA2-S per nmol/L increase in epinephrine in SS, and 98±32% in SR, ns), whereas the regression line for SS was shifted to the left (SS intercept 50.73±3.64%, SR 66.70±6.83%, p<0.03). This shift indicates that the effect of epinephrine on insulin resistance is enhanced in SS by 30 to 70%, depending on the absolute epinephrine value (the higher the epinephrine, the greater the percent reduction in HOMA2-S in SS compared to SR).
Higher heart rates correlated with insulin resistance in SR only. In contrast, higher plasma norepinephrine correlated with insulin resistance in SS only. Because levels of norepinephrine were about 20-fold higher than those of epinephrine in both groups, total catecholamines (epinephrine plus norepinephrine) also correlated with insulin resistance in SS only (r=0.30, p<0.04, not shown).
Within the SS group, the univariate correlates of HOMA2-S that cross-correlated with each other included epinephrine with norepinephrine, r=0.34, p<0.02, and mean arterial pressure with norepinephrine, r=−0.27, p=0.05. In SR, there were many cross-correlations between the univariate correlates of HOMA2-S. Hence, mean arterial pressure correlated with age (r=0.33, p<0.005), BMI (r=0.24, p<0.04), log aldosterone (r=0.60, p<0.00001), heart rate (r=0.44, p<0.0001) and epinephrine (r=0.25, p<0.04); heart rate with BMI (r=0.31, p<0.01), log aldosterone (r=0.50, p<0.00001) and epinephrine (r=0.47, p<0.0001); and log aldosterone with age (0.30, p<0.01) and epinephrine (r=0.29, p<0.02). Therefore, to assess the significant predictors of HOMA2-S in both groups, we employed multivariate regression. In the model for SR, only BMI and log aldosterone remained as significant predictors of HOMA2-S, accounting for half of its variability: HOMA2-S = 382.706 – 3.789*BMI – 79.792*log10Aldo, R2=0.508, F=37.153, p<0.0001. In SS, the significant predictors were mean arterial pressure, norepinephrine and epinephrine, with these catecholamines as separate variables having a greater predictive power than their sum as total catecholamines: HOMA2-S = 115.928 – 0.575*MAP 0.026*Norepi – 0.362*Epi, R2=0.310, F=6.8893, p<0.001. The explanatory power of this model (31% of the variability of HOMA2-S) was less than that for SR, suggesting that in SS there might be contributors to variability of insulin resistance that we did not study in our subjects.
We explored whether the characteristics of insulin resistance differ between SS and SR subjects, independent of BP (i.e., in normotensive and hypertensive subjects). We first confirmed observations by others that hypertensive patients are, in general, more insulin resistant than controls. Two thirds of patients attending hypertension clinics exhibit abnormalities in glucose metabolism11 and their reduced insulin sensitivity has been documented with euglycemic, hyperinsulinemic clamp26,27 and HOMA techniques14. In some28 but not all26 studies, normotensive offspring of hypertensive parents also had reduced peripheral glucose utilization. Because enhanced proximal tubular sodium reabsorption by insulin is preserved in hypertensive subjects27 and their offspring28, it has been suggested that insulin resistance may precede and predict the development of essential hypertension, a contention supported by epidemiological studies of large populations29. Insulin resistance increases with aging30 and our hypertensive subjects were somewhat older than our normotensive volunteers. However, it is unlikely that the 64% reduction in insulin sensitivity in the former can be solely attributed to an age difference of less than a decade.
A relationship between SSBP and insulin resistance has been documented in Dahl-S rats5,6 and confirmed in humans. Steady-state glucose-to-insulin ratio during an insulin and glucose infusion was twice as high in SS as in SR lean normotensive young subjects7, a magnitude of insulin resistance remarkably similar to that in our normotensive SS participants, in whom HOMA2-S was 47% of that of SR. In hypertensive patients classified into SS or SR by dietary8,10 or acute9 protocols, use of euglycemic clamp techniques confirmed these observations. In one study10, the relationship between SS and insulin resistance was independent of age, obesity, and hyperinsulinemia. We confirm this, since there were no differences in age, BMI or insulin levels between our SS and SR subjects, indicating that insulin resistance in SS is not due to these confounding factors. Increased odd ratios for developing insulin resistance, SSBP and lack of nocturnal dipping of BP (a feature of both SSBP and insulin resistance) have been linked to a polymorphism in the insulin-receptor-substrate type 1 (IRS-1) gene31, suggesting that the relationship between SSBP and insulin resistance may be genetically determined.
Our protocol, with salt loading and depletion, allowed us to assess the effect of salt balance on insulin resistance, a highly controversial issue. An early study with an euglycemic clamp in normal subjects showed that a high salt diet increased insulin resistance16. Two other studies in controls17,18, and one in diabetic subjects19 failed to detect any effect of dietary salt on insulin resistance, employing either clamp techniques or steady-state glucose concentration during an insulin suppression test. The opposite observation, i.e., increased insulin resistance by a low salt diet, has been made in small studies of normal subjects (17% reduction in insulin-mediated glucose disposal15), in a large group of normotensive participants of the HyperPath study (17% increase in HOMA insulin resistance index20) and in normotensive diabetic subjects studied with a euglycemic clamp (12% reduction in insulin sensitivity21). Stimulation of angiotensin II and catecholamines by a low-salt diet has been proposed as a mechanism for low salt diet-induced insulin resistance15. However, in the large HyperPath study the effect of low salt was independent of age, sex, BP, BMI, serum electrolytes, angiotensin II, PRA, aldosterone and catecholamines20. Only two studies assessed whether the effect of salt intake on insulin resistance depends on SSBP. In one, a high salt diet improved insulin sensitivity in SS normotensive men (not women) with a family history of hypertension22, consistent with worsening insulin resistance by a low salt intake, but in a sex- and SSBP-dependent manner. In the other, the conclusion was opposite, since glucose tolerance improved with salt restriction in healthy SS males whereas it worsened in SR23.
Our study cannot be compared with the ones above because all of them employed dietary changes in salt intake over several days, whereas ours utilized acute sodium balance manipulation and furosemide. However, our results agree with dietary studies showing worsened insulin resistance by salt-depletion, whereas we did not detect an effect of acute salt loading. Furthermore, we show worsening of insulin resistance by salt depletion only in the relatively insulin sensitive SR subjects (owing to 8% and 26% significant increases in glucose and insulin, respectively). In contrast, in the more severely insulin resistant SS group, salt depletion does not worsen it any further. This observation, if replicated in a dietary study without furosemide, would have obvious implications in terms of the controversy over recommended salt intake for hypertensive patients or normal populations. It would imply that salt restriction may be beneficial in SS subjects, reducing BP without worsening their already impaired metabolic status, whereas it may be detrimental in SR, not improving BP while inducing or worsening insulin resistance.
Many clinical and biochemical correlates of insulin resistance have been described. The association of obesity with insulin resistance in normotensive32 and hypertensive33 subjects may be due to accumulation of diacylglycerol in adipocytes and skeletal myocytes. Diacylglycerol-stimulated PKC isoforms serine-phosphorylate IRS-1 and IRS-2, inactivating the downstream PI3-kinase needed for translocation of GLUT 4 to the cell membrane34. A high salt intake may enhance this mechanism because it results in a 33% increase in plasma free fatty acids35. Indirect effects of obesity may include activation of sympathetic nervous system activity33 or inadequate suppression of aldosterone by salt36. The association of hypertension with insulin resistance is probably due to decreased skeletal muscle blood flow (chronic vasoconstriction and capillary rarefaction) with marked decrease in peripheral glucose uptake37 and impairment in the normal skeletal muscle vasodilatory effect of insulin38. Angiotensin II intereferes with insulin signaling via effects of nitric oxide39, NADPHoxidase40, and the servoregulatory negative feedback loop by SOCS (suppressor of cytokine signaling) proteins, which is shared by angiotensin II and insulin signaling and results in inhibiton of JAK/STAT pathways41. A relationship between PRA and insulin resistance was found in some studies in normotensive12 and hypertensive13 subjects but paradoxical improvement of insulin resistance by angiotensin II was shown by others42,43. Aldosterone inhibits insulin signaling in rodents44, isolated adipocytes, vascular smooth muscle45 and skeletal muscle46 via effects of inflammation and NADPH oxidase. Its actions are mediated by mineralocorticoid 44,46 or glucocorticoid47 receptors. Correlations between aldosterone and insulin resistance have been found in healthy subjects48, hypertension14 and heart failure49 and its plasma levels predict development of insulin resistance in populations50
Regarding a link between the sympathetic nervous system and insulin resistance, the most likely mechanism for norepinephrine is diminution of skeletal muscle glucose uptake owing to vasoconstriction51. For epinephrine, increased lipolysis, hepatic glucogenolysis and decreased peripheral utilization of glucose52,53 predominate, as supported by epinephrine-induced insulin resistance in normal men at subpressor doses54. Clinical studies have documented a relationship between heart rate and insulin resistance in normotensive and hypertensive subjects55. Insulin can stimulate the sympathetic nervous system but the sequence of events seems to be the opposite, i.e., autonomic alterations preceding the development of insulin resistance56,57.
Our finding of multiple univariate correlates of insulin resistance is consistent with the studies above. The novel and striking finding is that they are different between SR and SS, an observation not made in the only other publication that assessed this issue in direct fashion10. Univariate correlates of insulin resistance in our SR group (age, BMI, triglycerides, aldosterone, BP, heart rate and epinephrine) are consistent with those described in normotensive and hypertensive patients not phenotyped for SSBP. Because of cross-correlations between these variables, we resorted to multivariate modeling, which indicated that BMI and aldosterone, although inter-related, are independent predictors of insulin resistance in SR. The reason by which these variables were not predictors in SS cannot be ascertained from our study. Although blunted aldosterone responses to salt balance in SS may diminish the role of aldosterone in determining insulin resistance in this group, circulating levels of plasma aldosterone were only slightly, but not significantly lower, in SS than SR.
In SS, significant predictors included BP and both catecholamines (epinephrine and norepinephrine), consistent with previously observed stronger relationship between catecholamines and insulin resistance in SS than SR10. Their increased BP variability is consistent with reported relationships between insulin resistance and autonomic function56. Whereas epinephrine correlated with insulin resistance in both groups, supporting participation of its metabolic actions in all subjects, the left shift of this relationship in SS suggests the presence of additional factors in this group. The exclusive correlation of norepinephrine with insulin resistance in SS strongly suggests exaggerated peripheral vasoconstriction with diminished skeletal muscle glucose uptake as a major contributor in this group. Alternatively, different predictors in SR versus SS may simply reflect greater severity of insulin resistance or vasoconstriction in the latter (i.e., once established and severe, insulin resistance may depend on different mechanisms from those that operated at its onset). The decrease in the different magnitude of insulin resistance between groups as BP increased supports this possibility. Finally, it is noteworthy that we did not find relationships between PRA and insulin resistance in either group. Lack of this relationship in SS may be attributed to their blunted renin-angiotensin system but the reason in SR is less apparent, perhaps reflecting the conflicting effects of angiotensin II on insulin resistance quoted above.
Regardless of their reasons, different predictors suggest different underlying mechanisms for insulin resistance in SS versus SR. The obvious implication of this finding, if reproduced in larger populations with prolonged dietary protocols without use of furosemide, would be its applicability to selection of pharmacological therapy in hypertension. Whereas mineralocorticoid blockers could improve insulin resistance in SR, agents that suppress sympathetic activation should be more effective in SS.
In conclusion, our study confirms the presence of insulin resistance in hypertension and its greater severity in SS normotensive and hypertensive subjects compared to their SR counterparts. We also show differential effects of salt balance on and differential predictors of insulin resistance between groups, which may have implications for recommendations about salt restriction and pharmacological therapy in these two phenotypes.
We have shown that SS subjects have a high degree of insulin resistance compared to SR, and that acute salt loading and salt depletion do not further impair their insulin sensitivity. In contrast, SR are more insulin sensitive but when subjected to salt depletion, their insulin sensitivity is reduced to the level observed in SS. In addition, our modeling analyses suggest that the predictors of insulin resistance are different between groups; i.e., body mass index and aldosterone levels in SS versus mean arterial pressure and catecholamines in SR. We suggest that there are implications in terms of dietary recommendations for hypertensive patients or the population. Our findings predict that salt restriction will benefit SS by reducing their blood pressure without modifying their metabolic status, whereas it may render SR insulin-resistant without benefit on their blood pressure control. Also, different mechanisms for insulin resistance between these two groups may have implications for selection of pharmacological therapy in hypertension; e.g., use of agents that block the renin-angiotensin-aldosterone system versus those that act on the sympathetic nervous system.
First, the demonstration that the effects of salt depletion on insulin resistance are different in salt-sensitive and salt-resistant normotensive and hypertensive subjects. Second, the observation that the clinical and biochemical correlates of insulin resistance are also different between these two groups.
The different effects of salt depletion on insulin resistance have implications for health policy initiatives that advocate reduction of salt intake for the population at large and for all hypertensive patients. The different correlates of insulin resistance in salt-sensitive and salt-resistant hypertensive patients may be relevant for selection of antihypertensive therapy, to achieve simultaneous improvement of blood pressure and insulin sensitivity.
Salt depletion is beneficial for reducing blood pressure of salt sensitive subjects, which in the case of hypertension is the most prevalent phenotype. On the other hand, salt depletion will not benefit salt-resistant hypertensive patients, and may have adverse effects on their insulin sensitivity. Blockers of the renin-angiotensin-aldosterone system may be efficacious agents for improvement of insulin sensitivity in salt-resistant subjects whereas interventions that diminish sympathetic overactivity may be more efficacious in salt-sensitive ones.
Sources of Funding
This work was supported by grant MO1 RR00073 NCRR and by the Scott, Sherwood and Brindley Foundation.
Conflicts of interest:
Cheryl L. Laffer, Vanderbilt University School of Medicine, Nashville TN.
Fernando Elijovich, Vanderbilt University School of Medicine, Nashville TN.