Skip to main content
PMC home page
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2015 Nov 6.
Published in final edited form as: Hypertension. 2013 Feb 25;61(4):857–863. doi: 10.1161/HYPERTENSIONAHA.111.00712

Relationships Between Selected Gene Polymorphisms and Blood Pressure Sensitivity to Weight Loss in Elderly Persons With Hypertension

William J Kostis 1, Javier Cabrera 1, W Craig Hooper 1, Paul K Whelton 1, Mark A Espeland 1, Nora M Cosgrove 1, Jerry Q Cheng 1, Yingzi Deng 1, Christine De Staerck 1, Meredith Pyle 1, Nisa Maruthur 1, Ingrid Reyes 1, Cheryl AM Anderson 1, Jie Liu 1, John B Kostis 1
PMCID: PMC4636329  NIHMSID: NIHMS698523  PMID: 23438931

Abstract

Salt sensitivity, the heterogeneity in the response of blood pressure (BP) to alterations in sodium intake, has been studied extensively, whereas weight sensitivity, the heterogeneity in BP response to weight change, has received scant attention. We examined the relationship of 21 gene polymorphisms previously found to be associated with hypertension, diabetes mellitus, or obesity, with weight sensitivity in the Trial of Nonpharmacologic Interventions in the Elderly, where participants with hypertension were randomized to receive intensive dietary intervention of sodium reduction, weight loss, both, or attention control, whereas pharmacological therapy was kept constant. After correcting for multiplicity, we identified significant associations of 3 polymorphisms with weight sensitivity of systolic BP (rs4646994, rs2820037, and rs1800629) and 3 polymorphisms for diastolic BP (rs4646994, rs2820037, and rs5744292). A recursive partitioning algorithm selected the combination of rs4646994, rs1800629, rs1982073, and rs1800896 as the set associated with the highest weight sensitivity. Polymorphisms related to hypertension, obesity, and diabetes mellitus are associated with weight sensitivity of BP.

Keywords: blood pressure, genetics, obesity, sodium

Hypertension is a prevalent cardiovascular disease risk factor accounting for more death and disability than any other risk factor worldwide.1–3 Evidence relating salt intake and body weight to blood pressure (BP) derives from animal studies, physiology, epidemiology, and clinical trials.4–9 Salt sensitivity refers to heterogeneity among humans in their responses of BP to alterations in sodium and extracellular fluid volume status.10–15 Gene polymorphisms and ethnic variation account for part of the variability in salt sensitivity.16 Dietary sodium reduction and weight loss are effective in decreasing BP in patients with hypertension.9,17,18 Although salt sensitivity has been studied extensively, weight sensitivity, the heterogeneity in BP response to weight change, has received scant attention.19 Trial of Nonpharmacologic Interventions in the Elderly (TONE) tested whether sodium reduction or a weight loss program for obese participants would reduce the need for antihypertensive medication. In this study, compared with usual care, the need for antihypertensive therapy was lower after weight reduction or sodium reduction, with similar hazard ratios (0.60; 95% confidence interval, 0.45–0.80, and 0.64, 0.49–0.85, respectively).20 The purpose of this study was to examine whether polymorphisms related to hypertension, obesity, and diabetes mellitus are associated with weight sensitivity or salt sensitivity of BP.

Methods

TONE tested the hypothesis that sodium reduction in participants with hypertension or a weight loss program for those who were obese would reduce the need for antihypertensive medication. Using a 2×2 factorial design, obese (body mass index [BMI]>27.8 kg/m2 for men, 27.3 for women) participants were randomized to sodium reduction, weight loss, sodium reduction and weight loss combined, or attention control, whereas nonobese participants were randomized to sodium reduction or attention control for an average follow-up of 30 months. Men and women aged 60 to 80 years with systolic blood pressure (SBP) <145 mm Hg and diastolic blood pressure (DBP) <85 mm Hg on 1 or 2 antihypertensive medications were eligible to participate. Potential participants with serious cardiovascular disease, mental or physical illness, and insulin-dependent diabetes mellitus were excluded. The sodium-reduction goal was a 24-hour dietary sodium intake (by 24-hour urine sodium excretion) of 1800 mg (80 mmol) or less, and the weight loss goal was loss of 10 pounds or more. Participants randomized to attention control participated in meetings with lectures on topics not related to obesity, sodium intake, nutrition, or hypertension.

The intervention consisted of 3 phases (intensive, extended, and maintenance). The initial 4-month intensive phase provided participants with core knowledge and behavior skills, whereas the extended and maintenance phases focused on prevention of relapse. The data presented in this article pertain to the effect of the intensive intervention phase when participants received constant antihypertensive medication, while in the randomized dietary intervention. The first drug withdrawal visit for each patient occurred 76 to 104 days after the first group intervention session. Baseline values and the change in SBP, DBP, and BMI between baseline and the first drug withdrawal visit were calculated for each participant. The change in 24-hour urine sodium between baseline and the 9-month follow-up visit was also calculated (24-hour urine sodium was not measured at the first drug withdrawal visit).

Selection of Candidate Polymorphisms

We searched the literature for gene polymorphisms pertaining to hypertension, obesity, and diabetes mellitus. The most frequently published polymorphisms, and polymorphisms from the Wellcome Trust Consortium, were used to select 21 polymorphisms for study. Polymorphism data were collected in 2 stages: data for 7 polymorphisms were collected for the analysis performed for a previous angiotensin-converting enzyme (ACE) gene insertion/deletion (I/D) polymorphism,19 and data for an additional group of 16 polymorphisms were collected in 2011, including 2 polymorphisms (rs699 and rs5370) from the initial group, for a total of 21. Information on polymorphisms was available for 722 participants. Four hundred eleven of the 722 (56.9%) participants had at least 1 missing value, among which 219 (30.3%) had only 1 missing value. The polymorphisms and their reported associations with hypertension, diabetes mellitus, and obesity are shown in Table 1. Missing values were assigned by Bayesian imputation, controlling for the polymorphisms with available data and race.21

Table 1.

List of Selected Polymorphisms and Published Associations With Obesity, Diabetes Mellitus, and Hypertension

Genes Polymorphism Gene/Pathway Obesity Diabetes Mellitus Hypertension
rs5370 G/T Endothelin-1 X X
rs1800796 C/G Interleukin-6 X X X
rs7961152 A/C BCAT1* X
rs6997709 G/T KCNK9† X
rs4646994 I/D ACE X X X
rs5186 A/C AT2R1‡ X
rs2820037 A/T RYR2§ X X
rs5443 C/T GNB3∥ X X X
rs1937506 G/T Endothelin-1 X
rs1800629 G/A TNF-α# X X
rs4684847 C/T PPAR-γ X X X
rs11110912 C/G MYBPC1** X
rs1799983 G/T NOS3 X X X
rs1800872 A/C Interleukin-10 X
rs5744292 G/C Interleukin-18 X X X
rs699 A/C Angiotensinogen (AGT) X X
rs4961 G/T α-Adducin X X
rs1800795 C/G Interleukin 6 X X
rs1982073 C/T TGF-β1 X X X
rs1800896 A/G Interleukin-10 X X
rs187238 C/G Interleukin-18 X X
Open in a new tab
*

Branched-chain amino acid transaminase 1.

†

Potassium channel subfamily K member 9.

‡

Angiotensin II type 1 receptor (AGTR1) gene.

§

Ryanodine receptor 2.

∥

Guanine nucleotide binding protein beta polypeptide 3.

Endothelin-1/C198.

#

(G-308A) TNF promoter.

**

Myosin-binding protein C, slow-type.

Statistical Methodology

Linear Statistics

Modeling of SBP and DBP change was performed as a function of change in BMI or 24-hour sodium change and of the 21 polymorphisms chosen from the literature (Table 1), whereas controlling for participant characteristics and additional covariates.

Dependent Variables

The dependent variables were SBP_Change, the SBP at baseline minus the SBP at the first drug withdrawal visit; and DBP_Change, the DBP at baseline minus the DBP at the first drug withdrawal visit.

Independent Variables

The independent variables were subdivided into 5 groups:

  • (i).

    BMI-related covariates: Baseline_BMI and BMI_Change (from baseline to the first drug withdrawal visit).

  • (ii).

    Sodium-related covariates: Baseline 24-hour sodium and 24-hour_urine_sodium_Change (from baseline to the 9-month follow-up visit).

  • (iii).

    Genotype: 21 polymorphisms: The most important group of independent variables were the 21 polymorphisms denoted by I1,..., I21, where each Ii is a categorical variable taking values 1, 2, or 3 corresponding to wild type, heterozygous, and homozygous mutant, respectively. These independent variables were entered into the model either one at a time or all together.

  • (iv).

    Basic covariates: age and sex were included in all models.

  • (v).

    Investigator-selected covariates were 28 clinical covariates hypothesized to be related to SBP_Change, DBP_Change, and weight sensitivity. We ran a variable selection procedure, including variables listed under (i) to (v) using glmnet (a procedure designed to perform variable selection when the independent variables are correlated among themselves; Table S1 in the online-only Data Supplement).22

Data Subsets

The datasets analyzed were based on 6 randomized groups (a–f): (a) Obese participants, combined (sodium reduction and weight loss) intervention (n=112); (b) Obese participants, weight loss (n=112); (c) Obese participants, sodium reduction (n=99); (d) Obese participants, attention control (n=109); (e) Normal weight participants, sodium reduction (n=156); and (f) Normal weight participants, attention control (n=134). Seven subsets were defined as follows: (1) attention control (d and f); (2) weight loss (b); (3) sodium reduction (c and e); (4) combined intervention (a); (5) all sodium reduction (a, c, and e); (6) all weight loss (a and b); and (7) all participants (a–f).

Models

Six types of models were generated using combinations of the above groups:

  • Model 1: One polymorphism at a time+Baseline_BP+BMI_ Change+Baseline_BMI+24-hour_urine_sodium_Change+Baseline_24-hour_urine_sodium (examines weight and salt sensitivity)

  • Model 2: All polymorphisms+Baseline_BP+BMI_Change+ Baseline_BMI+24-hour_urine_sodium_Change+Baseline_ 24-hour_urine_sodium (examines weight and salt sensitivity)

  • Model 3: One polymorphism at a time+Baseline_BP+BMI_ Change+Baseline_BMI (examines weight sensitivity)

  • Model 4: All polymorphisms+Baseline_BP+BMI_Change+ Baseline_BMI (examines weight sensitivity)

  • Model 5: One polymorphism at a time+Baseline_BP+24-hour_ urine_sodium_Change+Baseline_24-hour_urine_sodium (examines salt sensitivity)

  • Model 6: All polymorphisms+Baseline_BP+24-hour_urine_sodium_Change+Baseline_24-hour_urine_sodium (examines salt sensitivity)

The 6 models were run with and without the investigator-selected covariates shown in independent variable list (v) above, yielding 12 linear models. Each model was run twice, once for SBP and once for DBP, for each of the 7 datasets (a total of 168 runs).

Recursive Partitioning Model

Although trees with splits based on polymorphisms have been used to predict the risk of coronary artery disease,23 the methodology used here is different. We built trees using only the polymorphisms to split the nodes with the other independent variables as the splitting criteria based on linear models 2, 4, and 6 described above.

Each tree yielded a partition of the data into m subsets that included individuals with a certain genotype, such that the relationship between SBP_Change (or DBP_Change) and BMI_Change (or 24-hour_urine_sodium_Change) follows a similar slope. Subsets of high interest are those with slopes significantly higher or lower than the average.

Polymorphisms were coded as: 1=Wild Type, 2=Heterozygous, or 3=Homozygous mutant. The partitions split the data into 2 groups in all 3 possible ways (1 versus 2 and 3, 2 versus 1 and 3, or 3 versus 1 and 2), and the partition that resulted in the biggest improvement in the R2 was chosen in each case.

Results

Of the 722 patients included in this study, 51.7% were men and the average age was 66±4.7 (median, 65; range, 60–80; interquartile range, 62–69). Five hundred thirty-eight (74.5%) were white and 181 (25.1%) were black. Two hundred forty-three were in the attention control group, 112 in the weight loss group, 255 in the sodium reduction group, and 112 in the combined intervention group. Baseline and first drug withdrawal visit BP, BMI, and urine sodium data are presented in Table S2. SBP_Change was –3.30±1.28 mm Hg in the weight loss group and –2.58±0.79 mm Hg in the sodium-reduction group. BMI_Change was –1.55±0.10 kg/m2 in the weight loss group and 24-hour_urine_sodium_Change was 49.15±4.00 mmol/24 hours in the sodium-reduction group (Table S2).

Linear Statistics

Selection of Clinical Variables

The glmnet procedure selected diabetes mellitus, current smoking, history of stroke >6 months before randomization, number of blocks walked per day, rate of physical activity during the year before being randomized, hours of moderate physical activity per week, hours of light physical activity per week, having a child with diabetes mellitus, and having a father with renal failure (Table S1).

Categorical Predictions Using Linear Models

Table 2 shows an array of linear model results, with 12 columns representing the 12 types of linear models for SBP. The rows are grouped in 7 blocks corresponding to the 7 datasets described in the methods. Each block shows the reference numbers of the polymorphisms and probability values of the associations that were statistically significant at the 0.1 level for the corresponding combination of column model type and row dataset. The data on the left of Table 2 are results of models adjusted for both 24-hour urine sodium and BMI, for 24-hour urine sodium only, and for BMI only, in each case without other covariates. Results of models considering both one polymorphism at a time, and all 21 polymorphisms are shown for each model type. Associations with probability values <0.1 are shown. The data on the right are results of models adjusting for sodium and BMI, sodium only, and BMI only, with the addition of the investigator-selected covariates. The results for weight sensitivity and salt sensitivity of DBP are shown in Table S3.

Table 2.

Associations of Gene Polymorphisms With the Relationship of SBP Change vs BMI and 24-Hour Sodium Change*

Na+Wt
 Wt Only
 Na Only
 Na+Wt+ Investigator-Selected
 Wt Only+ Investigator-Selected
 Na Only+ Investigator-Selected
Group Polymorphism One All One All One All Polymorphism One All One All One All
Group 1: Control rs5370 0.004 0.005 0.018 0.023 0.002 0.003 rs5370 0.059 0.065 0.039 0.046
rs1800796 0.018 0.020 0.017 0.018 rs7961152 0.034
rs1800795 0.093 rs1799983 0.025 0.076 0.044 0.020 0.056
rs1800896 0.085 0.096
rs7961152 0.086 0.086 0.018
rs11110912 0.090 0.073 0.097
rs1799983 0.058 0.083 0.051
Group 2: Wt
Only		 rs4646994 0.010 0.010 0.020 0.017 0.012 0.009 rs4646994 0.013 0.020 0.025 0.028 0.015 0.018
rs4961 0.092 0.098 0.096 rs5186 0.077 0.038 0.081
rs5186 0.057 0.027 0.059 rs2820037 0.039 0.036 0.040 0.014 0.042 0.033
rs2820037 0.029 0.010 0.030 0.005 0.032 0.010 rs6997709 0.091 0.095
rs6997709 0.088 0.098 0.093 0.096 rs1800872 0.091
rs1800872 0.096 0.077 0.095
Group 3: Na
Only		 rs4646994 0.086 0.096 0.048 0.053 0.089 0.097 rs4646994 0.050 0.063 0.025 0.032 0.052 0.065
rs5443 0.053 0.086 0.027 0.043 0.055 0.088 rs5443 0.058 0.098 0.029 0.048 0.060 0.100
rs1800896 0.072 rs7961152 0.091 0.095 0.094
rs4684847 0.024 0.083 0.047 0.025 0.084 rs4684847 0.021 0.039 0.022
rs6997709 0.001 0.004 0.006 0.028 0.001 0.004 rs6997709 <0.001 0.004 0.003 0.030 <0.001 0.004
Group 4:
Both Na+Wt		 rs2820037 0.016 0.032 0.018 0.038 0.014 0.032 rs2820037 0.027 0.047 0.024 0.038 0.028 0.061
rs11110912 0.061 0.092 rs5744292 0.091
rs1800629 0.013 0.003 0.015 0.009 0.014 0.005 rs11110912 0.061 0.028 0.062
rs1800872 0.066 0.077 0.078
Group 5:
All Na		 rs4646994 0.016 0.016 0.018 0.018 0.023 0.023 rs4646994 0.006 0.007 0.006 0.008 0.009 0.011
rs5443 0.035 0.042 0.040 0.056 0.036 0.041 rs699 0.063
rs2820037 0.096 rs1982073 0.089 0.090
rs4684847 0.004 0.003 0.007 0.005 0.004 0.003 rs2820037 0.096
rs1937506 0.022 0.056 0.047 0.028 0.066 rs7961152 0.085 0.043 0.048 0.098
rs6997709 0.003 0.057 0.009 0.082 0.003 0.048 rs4684847 0.007 0.010 0.013 0.015 0.008 0.011
rs11110912 0.083 rs11110912 0.068
rs6997709 0.006 0.039 0.017 0.08 0.006 0.034
Group 6:
All Wt		 rs2820037 0.002 0.001 0.001 0.001 0.002 0.001 rs2820037 0.002 0.001 0.001 <0.001 0.002 0.001
rs1937506 0.070 0.039 0.089 rs5744292 0.050 0.061
rs11110912 0.089 0.069 0.050 0.076 rs11110912 0.084 0.055 0.030 0.031 0.080 0.047
rs1800629 0.018 0.014 0.029 0.031 0.022 0.022 rs1800629 0.060 0.054 0.074 0.067 0.074 0.071
rs1800872 0.068 0.046 0.082 rs1800872 0.030 0.030 0.025 0.011 0.034 0.040
Group 7: All rs4646994 0.026 0.026 0.017 0.016 0.039 0.039 rs4646994 0.053 0.052 0.050 0.047 0.079 0.077
rs5370 0.028 0.036 0.042 0.052 0.021 0.024 rs699 0.093 0.048 0.056
rs2820037 0.072 0.047 0.057 0.048 0.074 0.051 rs5370 0.085 0.077 0.093 0.067 0.055
rs1800896 0.089 rs2820037 0.057 0.087 0.059 0.066
rs7961152 0.027 0.030 rs1800896 0.092 0.067 0.095
rs4684847 0.020 0.031 0.018 0.021 0.023 0.037 rs7961152 0.030 0.024 0.007 0.005 0.051 0.042
rs1937506 0.067 0.040 0.088 0.065 0.064 0.037 rs4684847 0.026 0.053 0.021 0.036 0.030 0.063
rs11110912 0.047 0.085 rs11110912 0.058 0.098
rs6997709 0.059 0.045 0.054 0.024 0.068 0.051
rs1800872 0.093 0.090 0.045 0.061 0.073
Open in a new tab

BMI indicates body mass index; and SBP, systolic blood pressure.

*

Changes between baseline and first drug withdrawal visit for SBP and BMI between baseline and 9-month visit for 24-hour sodium.

The table includes polymorphisms that were associated (P<0.1) with weight and salt sensitivity using linear statistics, as described in detail on page 6.

Controlling for Multiplicity

Predicted false discoveries and observed discoveries of associations of polymorphisms with SBP_Change or DBP_ Change in relation to BMI_Change and sodium reduction are shown in Tables S4 and S5. An issue with the analyses is the potential for spurious associations given the large number of comparisons performed. One approach to avoid this issue is to use Bonferroni correction to the α level and produce far more stringent criteria for significance. This would be appropriate, if the probability values were mutually independent. In this analysis, the polymorphisms are correlated, making the Bonferroni correction too strict. Instead, we calculated the number of false discoveries using our own adaptation of the Hochberg and Benjamini24 false discovery rate and applied it to our set of associations of polymorphisms. We simulated a set of hypothetical participants, where the genotype of each actual participant was linked to a randomly selected phenotype (all other variables). This process generated a null hypothesis in the sense that the genotype of each simulated participant had no relation to the phenotype, including the response variables SBP_Change and DBP_ Change. Therefore, any significant associations of polymorphisms with the other variables in this new random dataset would be false discoveries. We repeated this randomized simulation 100 times to estimate the number of expected false discoveries. We estimated the number of truly signifi-cant observed associations as the number of observed significant associations minus the number of false discoveries (at the median) at a given probability value. Tables S4 and S5 show the 50th and 80th percentiles of counts of false discoveries generated from 100 simulations of the random dataset for SBP (Table S4) and 100 simulations for DBP (Table S5). For both SBP and DBP, the third column shows the number of actual discoveries to compare with the false discoveries to estimate how many true discoveries (associations) are present in each case. Tables S4 and S5 were used in combination with Table 2 and Table S3 to determine which associations of polymorphisms with BP changes should be considered true discoveries and which are suspect of being false discoveries.

Summary of Data Obtained From Linear Statistical Analyses

In Table S6, we summarize the significant associations between the polymorphisms with SBP_Change and DBP_Change that were found by the linear models. The rule that we followed to select significant associations of polymorphisms with weight sensitivity and salt sensitivity was a combination of ranking the polymorphisms by the probability values of their associations with SBP_Change or DBP_Change, and correcting the list by eliminating the least significant polymorphisms according to the number of false discoveries.

In linear statistical analyses of the randomized groups of TONE, examining each polymorphism separately, we observed an association of weight sensitivity (BP change for a given change in BMI) with 5 polymorphisms with respect to SBP (rs1800629, rs4646994, rs5186, rs2820037, and rs1800872). Associations with DBP were observed for 6 polymorphisms (rs4646994, rs2820037, rs4684847, rs11110912, rs5744292, and rs4961). Among participants randomized to the combined intervention (both weight loss and sodium reduction), rs1800629 and rs2820037 were associated with weight sensitivity of SBP, and rs5744292 with weight sensitivity of DBP.

With respect to weight sensitivity and considering the simulation results of the random dataset (false discoveries) shown in Tables S4 and S5 and the results of the linear models (Table 2 and Table S3), 4 significant associations of polymorphisms with BP change (SBP or DBP) were identified. The rs4646994, an insertion deletion polymorphism of the angiotensin-converting enzyme and rs2820037, related to the ryanodine receptor 2, were significantly associated with weight sensitivity for both SBP and DBP. The rs1800629, related to tissue necrosis factor-α, was associated only with SBP and rs5744292, related to interleukin-18, only with DBP (Table 3). This table shows associations of polymorphisms and pathways of weight sensitivity and salt sensitivity observed in TONE, and corresponding data reported by previous investigators on salt sensitivity. Figure S1 shows box plots of weight sensitivity and salt sensitivity for selected polymorphisms.

Table 3.

Statistically Significant Categorical Predictions in TONE by Individual Polymorphisms/Genes/Pathways and Related Previously Reported Associations

Polymorphism Gene/Pathway SBP Weight Sensitivity DBP Weight Sensitivity SBP Salt Sensitivity DBP Salt Sensitivity Gene/Pathway Reference #
rs4684847 PPAR-γ 0.004 PPAR-γ 27
rs6997709 KCNK9 0.003 TGF-β and K+ 28
rs1800629 TNF-α 0.018 TNF-α and insulin resistance 29
rs2820037 RYR2 0.002 0.005 RYR2 and hypertension 30
rs4646994 ACE 0.01 0.004 ACE 31
rs5744292 IL-18 0.0001 0.005 IL-18 and hypertension and sudden death 32
rs7961152 BCAT1 0.03 BCAT1 and salt sensitivity 33
Open in a new tab

DBP indicates diastolic blood pressure; and SBP, systolic blood pressure.

P values correspond to a single polymorphism model adjusting for the basic covariates.

In the groups randomized to sodium reduction, salt sensitivity was associated with 4 polymorphisms with respect to SBP (rs6997709, rs4646994, rs5443, rs4684847), and with 3 polymorphisms with respect to DBP (rs5370, rs7961152, rs5744292). Polymorphisms rs6997709, related to potassium channel KCNK9, and rs4684847, related to PPAR-γ, had the strongest associations with salt sensitivity of SBP, and rs7961152, related to branched chain amino acid transaminase 1, and rs5744292, related to interleukin-18, with salt sensitivity of DBP (Table 3). Polymorphism rs5744292 was associated with both weight sensitivity and salt sensitivity of DBP.

Recursive Partitioning Model

The results of the recursive partitioning model for SBP and for DBP are shown in the online-only Data Supplemental Table S7. The algorithm produced 18 trees for SBP and 20 for DBP. Three of the 4 polymorphisms associated with weight sensitivity of SBP in the linear models (rs4646994, rs1982073, and rs1800896) were also selected by the recursive partitioning model as the combination with the highest SBP_Change per BMI_Change in the weight loss group. The slope increased with successive additions of polymorphisms from 3.51 mm Hg/kg per m2 in the initial division (rs4646994, 75 patients) to 5.55 mm Hg/kg per m2 (rs1800629, 49 patients), to 6.02 mm Hg/kg per m2 (rs1982073, 41 participants), to 6.51 mm Hg/kg per m2 (rs1800896, 31 patients; the online-only Data Supplemental Figure S2). The rs1800896 was also selected for weight sensitivity of DBP.

Discussion

This study shows significant associations of polymorphisms with the BP response to weight loss. After correction for multiplicity and adjustment for the expected false discoveries resulting from the large number of analyses, we identified 7 polymorphisms with significant associations for weight sensitivity. Two (rs2820037 and rs4646994) were associated with weight sensitivity of both SBP and DBP. In addition, one (rs1800629) was associated only with weight sensitivity of SBP, and another one (rs5744292) only with that of DBP.

Similar sets of polymorphisms were associated with weight sensitivity of SBP and DBP. Different associations or strengths of the associations for SBP and DBP with salt sensitivity were observed in the GenSalt study.25 It is possible that the polymorphisms that were identified as associated with weight or salt sensitivity could have been merely linked to causal variants, rather than causing variability in weight (and salt) sensitivity themselves.26

Although associations with the exact polymorphisms reported for salt sensitivity in TONE were not observed in the GenSalt cohort, polymorphisms affecting the same genes or pathways were observed for all 4 polymorphisms associated with weight and salt sensitivity as shown in Table 3.27–33 It is possible that the polymorphisms or causal variants linked to them affect BP and weight and salt sensitivity through a complex interplay of vasoconstrictive and vasodilating substances interacting with the autonomic nervous system. The recursive partitioning model identified that participants who had 2 of the 3 polymorphisms selected by the linear statistics had the highest weight sensitivity with a slope of 6.51.

Rhee and associates performed a controlled study of salt sensitivity in Korea, examining 36 polymorphisms previously reported to be associated with hypertension on 101 participants using ambulatory BP measurement.33 They examined 2 (rs6997709 and rs5744292) of the 4 polymorphisms significantly associated with salt sensitivity in TONE and found a significant association of one of them (rs5744292; odds ratio, 4.9; 95% confidence interval, 1.5–15.5; P=0.007).

The 2002 paper reported that I/D polymorphism of ACE was related to weight sensitivity.19 The present report expands on this finding by examining and adjusting for 20 additional polymorphisms and evaluating combinations of polymorphisms. One polymorphism examined in our study (rs5744292, related to interleukin-18) was associated with both weight sensitivity and salt sensitivity. Chen and associates have reported increased sodium sensitivity among individuals with the metabolic syndrome.15 We observed a significant association of the ACE I/D polymorphism (rs4646994) with weight sensitivity, but not with salt sensitivity. Zhang et al found that the ACE ID+II genotype interacted with daily salt intake and correlated with hypertension (P for interaction=0.047), and that BP was increased by high salt intake (P=0.005), whereas in the DD genotype it was not (P=0.257).31 This interaction was more prominent in the overweight group (P=0.039).

Limitations of the study include the fact that it is retrospective, and that the number of participants is small relative to the number of polymorphisms. Although the period under study was between randomization and the first drug withdrawal visit, information on 24-hour urine sodium was not available at the latter visit, as it was not collected during TONE. We substituted 24-hour urine sodium at the 9-month visit based on our finding that the change in 24-hour urine sodium between 9 and 18 months was small (45.2 versus 44.6 mmol/day, 1.3%),20 and that the average time interval between the first drug withdrawal visit and the 9-month visit was approximately 5 months (161±18.9 days; median, 161; interquartile range, 149–175).

Additional limitations are related to different temporal patterns of weight change and changes in physical activity among TONE participants.34 The missing data on individual polymorphism values is another limitation that we addressed with Bayesian imputation.

Strengths of the study include the precision of the clinical data collected during a randomized controlled observer-blind clinical trial, and that clinically relevant changes in body weight and salt intake were achieved. Also, these data from TONE are the only data describing weight sensitivity in hyper-tension. Additional strengths of this article are the statistical techniques used to overcome multiplicity and the recursive algorithm to examine combinations of individual polymorphisms, rather than only 1 polymorphism at a time.

In the last 50 years, great progress has been made in the identification of risk factors for cardiovascular disease, and in proving that control of these risk factors accrues significant clinical benefits from the points of view of morbid events as well as mortality. In addition, genomic studies have become more practical, faster, and less expensive.26 Thus, evaluation of weight sensitivity may be a way to identify individuals who may benefit more from weight loss as compared with other lifestyle interventions (eg, sodium reduction).

Perspectives

In conclusion, polymorphisms related to hypertension, obesity, and diabetes mellitus are associated with weight sensitivity of BP. Similar sets of polymorphisms are associated with weight sensitivity of SBP and DBP. This study addresses an issue that has received scant attention in the past. These results may help to clarify the pathophysiology of hypertension in some patient subsets and to identify individuals who are more likely to benefit from weight loss. Examination of additional polymorphisms or genome-wide association studies and study of the relationship of weight sensitivity to demographic, medication, and other patient characteristics may also be warranted.

Supplementary Material

Supplement

Novelty and Significance.

What Is New?

  • Although salt sensitivity has received considerable attention for >20 years, there is little information on the heterogeneity in the response of blood pressure to weight loss (weight sensitivity).

  • Polymorphisms related to hypertension, obesity, and diabetes mellitus are associated with weight sensitivity of blood pressure.

What Is Relevant?

  • These findings may be useful in clarifying the pathophysiology of hypertension in some patient subsets.

  • Evaluation of weight sensitivity may help identify individuals who may benefit more from weight loss as compared with other lifestyle interventions.

Summary

We identified significant associations of polymorphisms related to hypertension, obesity, and diabetes mellitus with weight sensitivity of blood pressure.

Acknowledgments

Sources of Funding

This study was supported, in part, by the National Institute on Aging, the National Heart, Lung, and Blood Institute, and the Robert Wood Johnson Foundation.

Footnotes

The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.111.00712/-/DC1.

Disclosures

None.

References

  • 1.Kearney PM, Whelton M, Reynolds K, Muntner P, Whelton PK, He J. Global burden of hypertension: analysis of worldwide data. Lancet. 2005;365:217–223. doi: 10.1016/S0140-6736(05)17741-1. [DOI] [PubMed] [Google Scholar]
  • 2.Mathers C, Stevens G, Mascarenhas M, World Health Organization . Global health risks: mortality and burden of disease attributable to selected major risks. World Health Organization; Geneva, Switzerland: 2009. [Google Scholar]
  • 3.Mathers C, Fat DM, Boerma JT, World Health Organization . The global burden of disease: 2004 update. World Health Organization; Geneva, Switzerland: 2008. [Google Scholar]
  • 4.Kostis JB, Davis BR, Cutler J, et al. Prevention of heart failure by antihypertensive drug treatment in older persons with isolated systolic hypertension. SHEP Cooperative Research Group. JAMA. 1997;278:212–216. [PubMed] [Google Scholar]
  • 5.He J, Whelton PK, Appel LJ, Charleston J, Klag MJ. Long-term effects of weight loss and dietary sodium reduction on incidence of hypertension. Hypertension. 2000;35:544–549. doi: 10.1161/01.hyp.35.2.544. [DOI] [PubMed] [Google Scholar]
  • 6.Sacks FM, Svetkey LP, Vollmer WM, Appel LJ, Bray GA, Harsha D, Obarzanek E, Conlin PR, Miller ER III, Simons-Morton DG, Karanja N, Lin PH. DASH-Sodium Collaborative Research Group. Effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet. DASH-Sodium Collaborative Research Group. N Engl J Med. 2001;344:3–10. doi: 10.1056/NEJM200101043440101. [DOI] [PubMed] [Google Scholar]
  • 7.Vollmer WM, Sacks FM, Ard J, Appel LJ, Bray GA, Simons-Morton DG, Conlin PR, Svetkey LP, Erlinger TP, Moore TJ, Karanja N. DASH-Sodium Trial Collaborative Research Group. Effects of diet and sodium intake on blood pressure: subgroup analysis of the DASH-sodium trial. Ann Intern Med. 2001;135:1019–1028. doi: 10.7326/0003-4819-135-12-200112180-00005. [DOI] [PubMed] [Google Scholar]
  • 8.Chobanian AV, Hill M. National Heart, Lung, and Blood Institute Workshop on Sodium and Blood Pressure: a critical review of current scientific evidence. Hypertension. 2000;35:858–863. doi: 10.1161/01.hyp.35.4.858. [DOI] [PubMed] [Google Scholar]
  • 9.Appel LJ, Moore TJ, Obarzanek E, Vollmer WM, Svetkey LP, Sacks FM, Bray GA, Vogt TM, Cutler JA, Windhauser MM, Lin PH, Karanja N. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med. 1997;336:1117–1124. doi: 10.1056/NEJM199704173361601. [DOI] [PubMed] [Google Scholar]
  • 10.Zemel MB, Sowers JR. Salt sensitivity and systemic hypertension in the elderly. Am J Cardiol. 1988;61:7H–12H. doi: 10.1016/0002-9149(88)91098-3. [DOI] [PubMed] [Google Scholar]
  • 11.Weinberger MH. Salt sensitive human hypertension. Endocr Res. 1991;17:43–51. doi: 10.1080/07435809109027188. [DOI] [PubMed] [Google Scholar]
  • 12.Hurwitz S, Fisher ND, Ferri C, Hopkins PN, Williams GH, Hollenberg NK. Controlled analysis of blood pressure sensitivity to sodium intake: interactions with hypertension type. J Hypertens. 2003;21:951–959. doi: 10.1097/00004872-200305000-00020. [DOI] [PubMed] [Google Scholar]
  • 13.He FJ, Markandu ND, MacGregor GA. Importance of the renin system for determining blood pressure fall with acute salt restriction in hypertensive and normotensive whites. Hypertension. 2001;38:321–325. doi: 10.1161/01.hyp.38.3.321. [DOI] [PubMed] [Google Scholar]
  • 14.Chen J. Sodium sensitivity of blood pressure in Chinese populations. Curr Hypertens Rep. 2010;12:127–134. doi: 10.1007/s11906-009-0088-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Chen J, Gu D, Huang J, Rao DC, Jaquish CE, Hixson JE, Chen CS, Chen J, Lu F, Hu D, Rice T, Kelly TN, Hamm LL, Whelton PK, He J. GenSalt Collaborative Research Group. Metabolic syndrome and salt sensitivity of blood pressure in non-diabetic people in China: a dietary intervention study. Lancet. 2009;373:829–835. doi: 10.1016/S0140-6736(09)60144-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Kelly TN, He J. Genomic epidemiology of blood pressure salt sensitivity. J Hypertens. 2012;30:861–873. doi: 10.1097/HJH.0b013e3283524949. [DOI] [PubMed] [Google Scholar]
  • 17.Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr, Jones DW, Materson BJ, Oparil S, Wright JT, Jr, Roccella EJ. National Heart, Lung, and Blood Institute Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; National High Blood Pressure Education Program Coordinating Committee. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289:2560–2572. doi: 10.1001/jama.289.19.2560. [DOI] [PubMed] [Google Scholar]
  • 18.Aronow WS, Fleg JL, Pepine CJ, et al. ACCF Task Force. ACCF/AHA 2011 expert consensus document on hypertension in the elderly: a report of the American College of Cardiology Foundation Task Force on Clinical Expert Consensus Documents. Circulation. 2011;123:2434–2506. doi: 10.1161/CIR.0b013e31821daaf6. [DOI] [PubMed] [Google Scholar]
  • 19.Kostis JB, Wilson AC, Hooper WC, Harrison KW, Philipp CS, Appel LJ, Espeland MA, Folmar S, Johnson KC. TONE Cooperative Research Group. Trial Of Nonpharmacologic interventions in the Elderly. Association of angiotensin-converting enzyme DD genotype with blood pressure sensitivity to weight loss. Am Heart J. 2002;144:625–629. doi: 10.1067/mhj.2002.123570. [DOI] [PubMed] [Google Scholar]
  • 20.Whelton PK, Appel LJ, Espeland MA, Applegate WB, Ettinger WH, Jr, Kostis JB, Kumanyika S, Lacy CR, Johnson KC, Folmar S, Cutler JA. Sodium reduction and weight loss in the treatment of hypertension in older persons: a randomized controlled trial of nonpharmacologic interventions in the elderly (TONE). TONE Collaborative Research Group. JAMA. 1998;279:839–846. doi: 10.1001/jama.279.11.839. [DOI] [PubMed] [Google Scholar]
  • 21.Su Y-S, Gelman A, Hill J, Yajima M. Multiple imputation with diagnostics (mi) in R: Opening windows into the black box. J Stat Softw. 2010;45:1–31. [Google Scholar]
  • 22.Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010;33:1–22. [PMC free article] [PubMed] [Google Scholar]
  • 23.Heng CK, He X, Saha N, Low PS, Demirci FY, Kamboh MI. Association of three lipoprotein lipase polymorphisms with coronary artery disease in Chinese and Asian Indians. Int J Cardiol. 2010;144:142–143. doi: 10.1016/j.ijcard.2008.12.121. [DOI] [PubMed] [Google Scholar]
  • 24.Hochberg Y, Benjamini Y. More powerful procedures for multiple significance testing. Stat Med. 1990;9:811–818. doi: 10.1002/sim.4780090710. [DOI] [PubMed] [Google Scholar]
  • 25.Zhao Q, Gu D, Hixson JE, Liu DP, Rao DC, Jaquish CE, Kelly TN, Lu F, Ma J, Mu J, Shimmin LC, Chen J, Mei H, Hamm LL, He J. Genetic Epidemiology Network of Salt Sensitivity Collaborative Research Group. Common variants in epithelial sodium channel genes contribute to salt sensitivity of blood pressure: The GenSalt study. Circ Cardiovasc Genet. 2011;4:375–380. doi: 10.1161/CIRCGENETICS.110.958629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.O'Donnell CJ, Nabel EG. Genomics of cardiovascular disease. N Engl J Med. 2011;365:2098–2109. doi: 10.1056/NEJMra1105239. [DOI] [PubMed] [Google Scholar]
  • 27.Zanchi A, Maillard M, Jornayvaz FR, Vinciguerra M, Deleaval P, Nussberger J, Burnier M, Pechere-Bertschi A. Effects of the peroxisome proliferator-activated receptor (PPAR)-gamma agonist pioglitazone on renal and hormonal responses to salt in diabetic and hypertensive individuals. Diabetologia. 2010;53:1568–1575. doi: 10.1007/s00125-010-1756-2. [DOI] [PubMed] [Google Scholar]
  • 28.Ying WZ, Aaron K, Wang PX, Sanders PW. Potassium inhibits dietary salt-induced transforming growth factor-beta production. Hypertension. 2009;54:1159–1163. doi: 10.1161/HYPERTENSIONAHA.109.138255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Shehata MF. Important genetic checkpoints for insulin resistance in salt-sensitive (S) Dahl rats. Cardiovasc Diabetol. 2008;7:19. doi: 10.1186/1475-2840-7-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Yamazaki D, Tabara Y, Kita S, et al. TRIC-A channels in vascular smooth muscle contribute to blood pressure maintenance. Cell Metab. 2011;14:231–241. doi: 10.1016/j.cmet.2011.05.011. [DOI] [PubMed] [Google Scholar]
  • 31.Zhang L, Miyaki K, Araki J, Song Y, Kimura T, Omae K, Muramatsu M. Interaction of angiotensin I-converting enzyme insertion-deletion polymorphism and daily salt intake influences hypertension in Japanese men. Hypertens Res. 2006;29:751–758. doi: 10.1291/hypres.29.751. [DOI] [PubMed] [Google Scholar]
  • 32.Hernesniemi JA, Karhunen PJ, Oksala N, Kähönen M, Levula M, Rontu R, Ilveskoski E, Kajander O, Goebeler S, Viiri LE, Hurme M, Lehtimäki T. Interleukin 18 gene promoter polymorphism: a link between hypertension and pre-hospital sudden cardiac death: the Helsinki Sudden Death Study. Eur Heart J. 2009;30:2939–2946. doi: 10.1093/eurheartj/ehp316. [DOI] [PubMed] [Google Scholar]
  • 33.Rhee MY, Yang SJ, Oh SW, Park Y, Kim CI, Park HK, Park SW, Park CY. Novel genetic variations associated with salt sensitivity in the Korean population. Hypertens Res. 2011;34:606–611. doi: 10.1038/hr.2010.278. [DOI] [PubMed] [Google Scholar]
  • 34.Kumanyika SK, Espeland MA, Bahnson JL, Bottom JB, Charleston JB, Folmar S, Wilson AC, Whelton PK. TONE Cooperative Research Group. Ethnic comparison of weight loss in the Trial of Nonpharmacologic Interventions in the Elderly. Obes Res. 2002;10:96–106. doi: 10.1038/oby.2002.16. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement
NIHMS698523-supplement-Supplement.pdf (277.9KB, pdf)

RESOURCES