| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology, Diabetes, and Hypertension (B.C., N.S.K., J.S.W., E.W.S., G.H.W.), Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts 02115; Cardiovascular Genetics Division (S.C.H.), University of Utah School of Medicine, Salt Lake City, Utah 84132; Department of Medicine (N.J.B., L.J.M.), Vanderbilt University Medical Center, Nashville, Tennessee 37232; and Department de Genetique (X.J.), Hôpital Européen Georges Pompidou, 75908 Paris, France
Address all correspondence and requests for reprints to: Gordon H. Williams, M.D., Division of Endocrinology, Diabetes, and Hypertension, Brigham and Women’s Hospital, 221 Longwood Avenue, Boston, Massachusetts 02115. E-mail: gwilliams{at}partners.org.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objectives: The objectives of this study were: 1) confirm bimodality of UFC, 2) assess whether UFC variations aggregate in hypertensive families, and 3) compare low-mode and high-mode UFC groups for distinguishing features.
Subjects/Setting: Subjects included 390 hypertensives and 166 normotensives from the general community.
Design/Interventions: Subjects had blood pressure and laboratory measurements on high- and low-salt diets. Familial aggregation was evaluated in 250 hypertensive siblings from 117 families.
Results: Hypertensives had higher UFC than normotensives (P < 0.001) and bimodal distribution of UFC (P < 0.0001). Analyses were controlled for gender and dietary sodium, which are confounding determinants of UFC. Mean low-mode UFC (33.8 ± 10.6 µg per 24 h) was similar to that of normotensives. The high mode, comprising 31.3% of hypertensives, had less change in mean arterial pressure between diets than the low mode (P = 0.01) without any other significant differences. Observed proportions of concordance and discordance for UFC mode differed significantly from that expected (P < 0.001). Observed concordance for the high mode was twice that expected, whereas for the low mode, it was similar to that expected by chance. Family membership explained a significant proportion of variance in UFC classification (P = 0.027). UFC mode of one sibling was a significant predictor of the UFC mode of the other sibling [odds ratio 6.6, 95% confidence interval (2.4–18.0), P < 0.001].
Conclusion: High-mode UFC is an intermediate phenotype of hypertension associated with salt resistance and a strong familial component supporting heritability.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The purpose of our study was 3-fold: 1) to confirm bimodality of UFC, 2) to assess whether UFC variations aggregate in hypertensive families, and 3) to compare low-mode (LM-UFC) and high-mode (HM-UFC) UFC groups for distinguishing features that may provide insight into potential mechanisms underlying the UFC variation and how it relates to hypertension.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Three hundred ninety hypertensives and 166 normotensives studied by the International Hypertensive Pathotype group are included in this report. Subjects were recruited from the general population of the Brigham and Women’s Hospital (Boston, MA) (n = 200), Hospital Broussais (Paris, France) (n = 124), University of Utah (Salt Lake City, UT) (n = 182) and Vanderbilt University (Nashville, TN) (n = 50). The institutional review board of each institution approved the study. All subjects gave written informed consent before enrollment and underwent a standardized protocol. Some characteristics of subsets of this population have been reported previously (10, 12, 13, 14). However, the present analyses are original.
Two hundred seventy-eight of the 390 hypertensives in this study were new subjects, not part of the study done by Litchfield et al. (10). Data from this new cohort were analyzed for reproducibility of previous observations (10). The total cohort of 390 hypertensives, including 250 members from 117 multisibling families (102 sibling pairs; 14 trios; one family with four siblings) was analyzed for familial aggregation.
All subjects had a screening history and physical and laboratory examination. Subjects with known or suspected secondary hypertension, diabetes, coronary artery disease, stroke, overt renal insufficiency (serum creatinine > 1.5 mg/dl), psychiatric illness, other significant medical illness except obesity, current oral contraceptive use, current tobacco/illicit drug use or alcohol intake greater than 12 oz/wk were excluded. Subjects with abnormal electrolytes or thyroid/liver function tests or electrocardiographic evidence of heart block, ischemia, or prior coronary events at the screening exam were excluded. Subjects were between ages 18 and 65 yr. Race was self-defined.
Hypertension was defined as seated diastolic blood pressure (DBP) 100 mm Hg or greater off antihypertensive medication, 90 mm Hg or greater with one or more medications, or treatment with two or more medications. Subjects requiring more than four antihypertensive medications were excluded. Normotensives, in addition to having blood pressure (BP) less than 140/90 mm Hg, reported no first-degree relatives diagnosed with hypertension before age 60 yr. Subjects taking an angiotensin converting enzyme inhibitor, angiotensin receptor blocker, or mineralocorticoid receptor antagonist were transitioned to amlodipine 3 months before study to minimize interference with assessment of the renin-angiotensin-aldosterone system. If needed, hydrochlorothiazide was added to control BP. All antihypertensive medications were discontinued 2 wk before study.
Subjects completed two dietary phases: 5–7 d of high-sodium (200 mmol/d) and 7 d of low-sodium (10 mmol/d) diets. Each diet was isocaloric, contained 100 mmol/d potassium and 20 mmol/d calcium, and was caffeine and alcohol free. On the final day of each diet, subjects were admitted to the General Clinical Research Center and remained fasting and supine overnight. A 24-h urine collection was used to confirm sodium balance (
150 mmol for high sodium and 
30 mmol for low sodium) and for creatinine and UFC measurements. Hemodynamic and laboratory assessments, including plasma cortisol, were made in the morning. BP was measured at 5-min intervals using an automated device (DINAMAP; Critikon, Tampa, FL). Three consecutive readings were averaged for analysis.
Mean arterial pressure (MAP) was calculated as 1/3 [systolic BP (SBP) – DBP] + DBP. BP salt sensitivity was assessed as change in MAP between the two diets. BP response to angiotensin II was assessed as the change in BP from baseline to after infusion of 3 ng/kg·min of angiotensin II for 55 min. Effective renal plasma flow, as p-aminohippuric acid clearance, was calculated from steady-state plasma p-aminohippuric acid clearance concentrations as previously described (14, 15). Renal vascular resistance (RVR) was calculated as the ratio of MAP to effective renal plasma flow. The homeostasis model assessment (HOMA) index was calculated as [fasting plasma glucose (millimoles per liter) x fasting plasma insulin (milliinternational units per liter)]/22.5.
Plasma renin activity (PRA) was measured in low-sodium balance after 1–2 h in the standing position. Hypertensives were classified as low-renin if PRA was less than 2.5 ng/ml/hr (<0.69 ng/liter per second) or normal renin. The normal-renin group was subdivided into nonmodulators if plasma aldosterone response was 15 ng/d or less (416.2 pmol/liter) above baseline to infused angiotensin II [3 ng/kg·min for 55 min] and modulators (all others not included in the prior two groups) as previously described (12, 13).
Laboratory procedures
UFC was measured by Coat-A-Count RIA (Diagnostics Products Corp., Los Angeles, CA), with sensitivity of 0.2 µg/dl (5.5 nmol/liter) and precision of 4–6.4%. Serum cortisol was measured by Access Cortisol assay (Beckman Coulter, Chaska, MN), with sensitivity of 0.4 µg/dl (11 nmol/liter) and precision of 6.4–7.9%. UFC and urinary cortisone were measured by HPLC in a subset. Details of most other laboratory procedures have been described previously (13, 14, 16).
Statistical analysis
Data are reported as mean ± SD for continuous variables and percentages for discrete variables unless otherwise specified. Statistical significance was indicated by a P < 0.05. Statistical analyses were performed using SPSS (version 14.0; SPSS, Chicago, IL) and SAS (version 8.2; SAS Institute, Cary, NC) software packages. Comparison of groups (hypertensives with normotensives; HM-UFC to LM-UFC) was done using mixed effect linear regression (PROC MIXED) clustered by family to account for nonindependence of related subjects. Data were analyzed for determinants of UFC by univariate and multivariate methods. UFC data were analyzed for bimodality using maximum likelihood analysis, which tests whether the distribution is unimodal vs. two normal distributions, using gender-adjusted residual UFC values. The mean of the lower mode of the bimodal distribution plus 2 SD was used as the cutoff to stratify hypertensives into low- and high-mode groups.
We tested the hypothesis that UFC variations clustered within hypertensive families using two approaches, without assuming any particular ordering of siblings within families. The first approach tested whether the observed familial aggregation for either UFC mode exceeded the aggregation occurring by chance. The expected concordance/discordance for UFC mode in the general population was estimated using the proportion of low-/high-mode individuals in the unrestricted database of 390 hypertensives. A 
2 analysis was used to test the significance of the excess observed aggregation. The ratio of observed to expected concordance, which would be one if the trait were randomly distributed, was determined. The analysis was repeated using four methods for weighting families: 1) weight based on number of members/sibling set, 2) equal weight/sibling set irrespective of number, 3) each sibling set considered a single unit and all siblings/unit having to fall within the same UFC mode to be considered concordant, and 4) sibling sets with more than two members considered, using all possible combinations, as multiple sibling pairs (three pairs/trio; six pairs from the four sibling family) with each pair receiving equal weight. The second approach to assess for familial influence on UFC used mixed-models logistic regression (SAS NLMIXED) with UFC mode as the binary outcome and family membership as the random effect. In addition to the above approaches, using multivariate logistic regression, families with only two siblings (n = 102) were analyzed to assess whether, within sibling pairs, the UFC mode of one sibling predicts that of the other.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Characteristics of the new cohort of hypertensives and the normotensives are shown in Table 1
. Both groups achieved target sodium balance per the study protocol. UFC was higher on high-salt diet (Fig. 1). Gender significantly modified UFC in hypertensives on both diets (Fig. 1). In normotensives, the gender effect was observed on high-salt diet (Fig. 1). Hypertensives had higher UFC than normotensives, even after controlling for gender differences. Although normotensives and hypertensives differed with respect to age and body mass index (BMI), we found no relationship between UFC and these variables. There was a strong correlation between high-sodium and low-sodium UFC (r = 0.51, P < 0.001). Serum cortisol was slightly lower in hypertensives than normotensives on low-sodium diet (P = 0.02) but similar on high-sodium diet.
|
|
By maximum likelihood analysis, gender-adjusted UFC data for the new cohort of hypertensives fit a bimodal distribution significantly better than a unimodal distribution on both diets (P < 0.0001). Having replicated the findings of the study by Litchfield et al. (10) in a new cohort, we pooled all hypertensives, i.e. the new cohort (278 subjects) and 112 subjects from the previous study (10). Analysis of this combined cohort also revealed bimodality of UFC. For the combined cohort, mean overall UFC estimated by maximum likelihood analysis for the two normal distributions were 33.5 and 63.3 µg per 24 h with SD of 10.2 and 25.6 and proportions of 0.57 and 0.43, respectively. The lower mean + 2 SD was used as the cutoff to define low-mode and high-mode groups for further analyses. This approach was taken to minimize the potential contamination of individuals with low-mode UFC in the higher mode subgroup.
Analysis of the combined/total cohort of hypertensives
Comparison of the combined cohort to the normotensives yielded results similar to those in Table 1
for the new cohort except gender distribution was now similar. The combined cohort also showed similar effects of gender and dietary sodium and bimodality as described above.
Familial aggregation of UFC
The observed proportions of hypertensive siblings concordant/discordant for UFC mode differed significantly (P < 0.001) from that expected by chance (Fig. 2). Weighting families based on the number of siblings contributing to the analysis, the observed concordance for HM-UFC (19.4%) was 2.03 times that expected (9.6%), whereas the concordance for LM-UFC (49.9%) was similar to that expected by chance (47.6%) (Fig. 2). Observed discordance for UFC mode was less than expected (30.7 vs. 42.8%). Repeating the analysis using the other three methods for weighting families, as outlined in Statistical analysis, yielded similar results, with concordance for HM-UFC remaining almost 2-fold higher than expected, indicating familial aggregation, whereas concordance for LM-UFC was similar to that expected for a randomly distributed trait.
|
Comparison of high mode vs. low mode
LM-UFC and HM-UFC groups represented 68.7 and 31.3% of the hypertensives, respectively. Table 2 presents the characteristics of the two groups. Age, gender, and race distributions were similar. HM-UFC had slightly higher low-salt serum cortisol, but high-salt levels were similar (P = 0.08), thus making the relevance of this finding unclear. Mean UFC of the low mode was similar to that of normotensives.
|
BMI, cholesterol indices, and HOMA index were similar in both groups. Interestingly however, although the differences were not statistically significant, HM-UFC had slightly lower mean BMI, triglyceride, and HOMA index.
Mineralocorticoid characteristics
Serum potassium and aldosterone levels were similar in both groups. HM-UFC had higher PRA. Both groups achieved target sodium balance.
Hemodynamic characteristics
To confirm the previously reported relationship between UFC and salt sensitivity (10), we assessed BP response to dietary sodium. HM-UFC had significantly less change in MAP (Fig. 3), compared with LM-UFC (P = 0.01). HM-UFC had higher SBP on a low-sodium diet (P < 0.0001) and significantly less change in SBP between diets (P = 0.002), also indicating more salt resistance. Baseline DBP and the change in DBP between diets (P = 0.06) were similar. However, the HM-UFC group showed a slightly greater DBP response to angiotensin II on low-salt diet (P = 0.01). The SBP response to angiotensin II was similar in both groups.
|
There were no significant differences in renal blood flow or RVR in a subset. High- and low-salt RVR data were available for 233 (177 LM-UFC; 56 HM-UFC) and 154 subjects (130 LM-UFC; 24 HM-UFC), respectively. Urinary cortisone and UFC measured by HPLC in a subset on high-salt (20 LM-UFC; 12 HM-UFC) and low-salt diets (21 LM-UFC; 10 HM-UFC) showed that the HM-UFC group had higher urinary cortisone on both diets (P = 0.001 and 0.027) and higher UFC (Fig. 4), which was significant on low-salt diet (P < 0.001; high salt, P = 0. 057). Cortisol to cortisone ratio was similar in both groups on both diets (P = 0.08 and 0.10). Low-renin and nonmodulating hypertensive phenotypes, which are known to be salt sensitive, were underrepresented in the high mode (63.1% modulators, 17.2% nonmodulators, and 19.7% low renin) as opposed to 50.4% of the low mode with salt-sensitive low-renin or nonmodulating forms of hypertension. The fewer low-renin hypertensives likely also accounts for the higher PRA in the high mode.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Essential hypertension is determined by the interplay of multiple genetic and environmental factors, which may obscure the association between a gene and this complex trait. Classifying hypertensives into more homogenous groups (intermediate phenotypes), based on a common phenotypic trait reduces this heterogeneity. We previously used this approach successfully to identify genetic underpinnings of two forms of salt-sensitive hypertension (12, 13). Whereas there has been substantial progress in identifying genetic factors in salt-sensitive hypertension, little is known about genetic variants responsible for salt-resistant hypertension.
Genetic variations in corticosteroid metabolism/function are known to cause hypertension in conditions such as apparent mineralocorticoid excess syndrome (1) and the glucocorticoid resistance syndromes (17). Whereas there are reports of abnormalities in cortisol metabolism (6) and increased tissue sensitivity to cortisol in essential hypertension (5), evidence as to whether these are genetically determined is limited. The Four Corners study (9) reported higher serum cortisol and increased frequency of a particular allele of the glucocorticoid receptor gene in young adults, with relatively higher personal BP and hypertensive parents, compared with peers with lower BP and nonhypertensive parents. A subsequent study demonstrated that young adults with higher BP and hypertensive parents had enhanced glucocorticoid sensitivity, increased cortisol secretion, and impaired conversion of cortisol to inactive metabolites (7). These findings suggest that genetically mediated abnormalities in cortisol may contribute to familial predisposition for hypertension, but they have not been reproduced in a hypertensive population.
What biological mechanisms could explain our observation of high UFC in a subset of hypertensives and is there a difference in the mechanism of hypertension in this subset? First, could the HM-UFC group have Cushing’s syndrome? There are several factors that argue against this, besides the fact that these subjects had no history or clinical evidence of this diagnosis. Their metabolic profiles including BMI, lower triglycerides, and less insulin resistance seem inconsistent with what would be expected with Cushing’s. The UFC levels were also within the high-normal range for several in this group and majority had UFC less than 100 µg per 24 h. Besides, the high mode constituted 31% of our hypertensives. Given the relative infrequency of Cushing’s syndrome in the general population, it would be highly unlikely that such a large proportion of hypertensives could have hitherto undiagnosed subclinical Cushing’s syndrome. However, dexamethasone suppression tests were not performed. Conversely, our results also have implications for using UFC to screen hypertensives for subclinical Cushing’s syndrome because care must be taken to avoid wrongly assigning this diagnosis based on a randomly collected UFC level without taking into account dietary sodium, gender, and the overall clinical picture.
Could the UFC elevation be due to abnormal 11-ß hydroxysteroid dehydrogenase (11-ß HSD) activity from either a genetic defect or inhibitors of the enzyme? 11-ß HSD exists as two isoforms: type I, which mostly converts cortisone to cortisol, and type II converting cortisol to cortisone in the kidney. Prior reports have suggested a role for 11-ß HSD type 2 abnormality in essential hypertension (6, 18). Such an abnormality would impair conversion of cortisol to cortisone, resulting in excess cortisol activating the mineralocorticoid receptors, causing renal sodium and fluid retention leading to hypertension. In this setting, one would anticipate salt-sensitive hypertension, but our HM-UFC was salt resistant. Besides, urinary cortisol to cortisone ratio analyzed in a subset was preserved in the HM-UFC group, suggesting that both isoforms of the enzyme are functioning normally. The absence of elevated aldosterone or potassium in the HM-UFC group also indicates that there is no activation of the aldosterone axis.
Could there be an abnormality of the glucocorticoid receptor (GR) in the HM-UFC group? There is some prior evidence suggesting GR variation/abnormality in essential hypertension (9, 19). Glucocorticoid resistance syndromes caused by genetic GR alterations resulting in impaired cortisol binding, decreased target-tissue sensitivity to cortisol, and compensatory increase in ACTH and hence cortisol but without physiologic manifestations of cortisol excess, have been described in New World primates (20) and humans (17). Relevant data to further analyze this possibility were not available, but this is certainly an intriguing mechanism warranting further evaluation.
Alternatively, could cross-reactivity of another substance with the cortisol assay be causing falsely elevated UFC values? The UFC assay we used has very little documented cross-reactivity (<1%) for compounds such as aldosterone, corticosterone, cortisone, 11-deoxycorticosterone, and tetrahydrocortisol. It cross-reacts with prednisolone (76%) and 11-deoxycortisol (11%), but it is unlikely that these compounds are driving our findings. In addition, UFC measured in a subset by HPLC was also higher in the high mode. Whereas we cannot completely exclude cross-reaction by another unidentified substance, it is unlikely.
Finally, could the UFC elevation reflect an altered stress response? Is it mediated by the sympathetic nervous system? Data were inadequate in our current data set to assess differences in catecholamines. HM-UFC showed slightly greater DBP response to angiotensin II infusion, suggesting higher pressor responsiveness, but the clinical significance of this small difference in DBP response is unclear.
The mechanisms whereby cortisol raises BP are not clearly understood. Previous studies have shown that cortisol-induced hypertension is accompanied by sodium retention and volume expansion, but this is not the primary mechanism causing hypertension (21). Whereas cortisol increases cardiac output, a rise in cardiac output is not essential for the rise in BP (22). Previous studies have also shown that cortisol-induced hypertension is not accompanied by increases in vasopressor hormones and sympathetic activity (23, 24, 25). There is evidence suggesting that glucocorticoids modulate vascular function (5, 24) by increasing vascular reactivity (5) and vascular pressor responsiveness (24, 26). Our HM-UFC group showed slightly increased DBP response to angiotensin II, but this effect seems too small to solely account for the hypertension in this group. A role for increased RVR has been suggested (27) but our analysis did not show a difference in RVR or renal blood flow between subsets of the two modes. Abnormalities of the nitric oxide system have been suggested as a potential mechanism (28, 29), but details remain unclear and need further investigation.
Our findings should be interpreted in the context of the study design. They suggest, but do not prove, the existence of genetic influences on UFC in hypertensives. They confirm an association between HM-UFC and a relatively salt-resistant subgroup of essential hypertensives, but a causal relationship and the underlying mechanisms remain to be established. Hypertensives in our database had mild/moderate hypertension. Therefore, results may not be applicable to severe hypertensives. Our subjects were under age 66 yr. Hence, results may not apply to older individuals. The strengths of this study are the large sample size, ability to control for relevant confounders, carefully controlled experimental conditions, and the fact that hypertensives were off medications when studied, thereby eliminating the interaction of drug effects.
In summary, this study has established that familial determinants independently contribute significantly to UFC levels in hypertensives, and a subset of relatively salt-resistant hypertensives, with predominantly normal PRA levels, has higher UFC. Genetic causes are probable but not definite. Further investigation is required to better characterize this intermediate phenotype, determine the genetic underpinnings, and eventually develop a specific therapeutic approach for this subgroup of hypertensives.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online January 30, 2007
Abbreviations: BMI, Body mass index; BP, blood pressure; DBP, diastolic blood pressure; GR, glucocorticoid receptor; HM-UFC, high-mode UFC; HOMA, homeostasis model assessment; 11-ß HSD, 11-ß hydroxysteroid dehydrogenase; LM-UFC, low-mode UFC; PRA, plasma renin activity; RVR, renal vascular resistance; SBP, systolic BP; UFC, urinary free cortisol.
Received September 25, 2006.
Accepted January 24, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-adducin polymorphism. Hypertension 39:191–196This article has been cited by other articles:
 |
|
 |
|
  G. Reimondo, S. Bovio, B. Allasino, S. De Francia, B. Zaggia, I. Micossi, A. Termine, F. De Martino, P. Paccotti, F. Di Carlo, et al. The combined low-dose dexamethasone suppression corticotropin-releasing hormone test as a tool to rule out Cushing's syndrome Eur. J. Endocrinol., November 1, 2008; 159(5): 569 - 576. [Abstract] [Full Text] [PDF]
|
|
Read all eLetters
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
