Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Jul 1.
Published in final edited form as: J Hypertens. 2016 Jul;34(7):1266–1272. doi: 10.1097/HJH.0000000000000936

SERUM FIBROBLAST GROWTH FACTOR-23 AND INCIDENT HYPERTENSION: THE ATHEROSCLEROSIS RISK IN COMMUNITIES STUDY

Amber L FYFE-JOHSNON 1, Alvaro ALONSO 1, Elizabeth SELVIN 2,3, Julie K BOWER 4, James S PANKOW 1, Sunil K AGARWAL 5, Pamela L LUTSEY 1
PMCID: PMC4950510  NIHMSID: NIHMS800563  PMID: 27100793

Abstract

Elevated serum fibroblast growth factor-23 (FGF23), an endogenous hormone, is associated with disturbed mineral homeostasis, cardiovascular disease, and chronic kidney disease. It is unclear whether FGF23 impacts the development of incident hypertension. We investigated the association of serum FGF23 measured at baseline (1990–92) with incident hypertension at two follow-up visits (1993–95 and 1996–98) in 7,948 middle-aged men and women without hypertension at baseline participating in the Atherosclerosis Risk in Communities Study. Incident hypertension was determined by measured blood pressure (DBP ≥ 90 mm Hg, or SBP ≥ 140 mm Hg) and/or self-reported hypertension medication use at follow-up exams. Complementary log-log models that accounted for interval censoring were used to model the association between FGF23 and incident hypertension. During a median follow-up of 5.9 years, 27% (2,152/7,948) participants developed hypertension. A nonlinear association between serum FGF23 and incident hypertension was observed; only persons in the highest decile of serum FGF23 had an increased risk of incident hypertension. After adjustment for demographics, behaviors, and adiposity the hazard ratio for incident hypertension was 1.24 (95% CI: 1.11, 1.39) for the highest decile of FGF23 compared to the lowest quintile. The association was further attenuated in the final model after adjusting for renal function (hazard ratio: 1.21, 95% CI: 1.08, 1.35). In conclusion, high levels (≥ 60.6 pg/mL) of FGF23 are associated with a modestly increased risk of incident hypertension in the general population, independent of kidney function.

Keywords: Fibroblast growth factors, fibroblast growth factor-23, hypertension, Atherosclerosis Risk in Communities Study, risk factors

INTRODUCTION

Approximately 78 million Americans adults have hypertension1; the age-adjusted prevalence approaches 65–75% among persons aged 65 years or older.2,3 Hypertension frequently underlies cardiovascular disease; evidence suggests that hypertension is present in 69% of adults with incident myocardial infarction, 77% with incident stroke, and 74% diagnosed with incident heart failure.1,3 Development of hypertension occurs in conjunction with, (i) the inability of the renal system to maintain mineral homeostasis, and (ii) changes in cardiac and vascular structure.4

Fibroblast growth factor-23 (FGF23) is a protein produced by osteocytes and osteoblasts. It is principally involved in kidney function and mineral metabolism;5,6 elevated levels have been linked to cardiovascular morbidity and mortality in both the general population7–9 and those with chronic kidney disease.10–12 It is unclear whether FGF23 impacts the development of hypertension, and if so whether the biological mechanism involves mineral metabolism, changes to cardiac or vascular structure, or another explanation. Recent work demonstrates that elevated FGF23 is associated with mineral homeostasis disruption, plasma volume, and hypertension in animal models.13,14 Importantly, in individuals with normal kidney function, FGF23 is a risk factor for mortality, heart failure, and cardiovascular disease events.15,16 Therefore, it appears that FGF23 may have cardiovascular consequences in conjunction with and independent of mineral metabolism.

The objective of the current study was to evaluate the association between FGF23 and incident hypertension in the prospective, community-based Atherosclerosis Risk in Communities (ARIC) Study. We hypothesized that elevated serum FGF23 would be positively associated with risk of incident hypertension independent of traditional cardiovascular (CVD) risk factors in the ARIC cohort.

METHODS

Study Population

The Atherosclerosis Risk in Communities (ARIC) Study is a longitudinal prospective cohort study initiated in 1987–89. Originally, 15,792 men and women ages 45–64 years of age were recruited from 4 U.S. communities: Minneapolis suburbs, Minnesota; Jackson, Mississippi; Washington County, Maryland; and Forsyth County, North Carolina.17 Five follow-up visits were conducted: visit 2 (1990–92), visit 3 (1993–95), visit 4 (1996–98), and visit 5 (2011–13). FGF23 was first measured at visit 2, thus visit 2 is baseline for the present analysis. Of the 14,348 participants attending visit 2, we excluded participants with prevalent hypertension (defined as measured systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg and/or self-reported antihypertensive medication use) at visit 2 (n=5,163), individuals who were neither African American nor white, and African Americans from the Minnesota and Maryland centers due to small sample sizes (n=91), individuals with FGF23 values >200 pg/mL (n=12) or missing FGF23 values (n=751), and those missing hypertension classification at visits 3 and 4 (n=383). Our final sample size was 7,948 observations. No exclusions were made for missing covariates. Written informed consent was collected from all study participants; all affiliated institutional review boards approved the study protocol.

FGF23 and other covariates

FGF23

All participants were instructed to fast overnight prior to blood specimen collection at study visit 2 (1990–92). Blood samples were processed, shipped, and stored at −70°C until laboratory assays were performed.17 In 2012–2013, intact serum FGF23 was measured in stored specimens collected at ARIC visit 2 by enzyme-linked immunosorbent assay (Kainos Laboratories, Inc., Tokyo, Japan). Serum samples were incubated in microtiter plates previously coated with a capture antibody. A detection antibody was then added to form a sandwich complex between serum FGF23 that had bound to the detection antibody on the plate. The microtiter plate was then washed and incubated with a tetramethylbenzidine substrate solution. Finally, a spectrophotometric plate reader determined the optical density of each well and sample intact FGF23 concentrations were calculated from a standard curve. The inter-assay (process) coefficient of variation for FGF23 based on ARIC blind duplicate samples (split-specimens collected at the time of blood draw) was 16.6% and the coefficient of variation from internal laboratory quality control samples was 8.8%; internal lab quality control was based on a standard blood pool which had a serum FGF23 concentration of 41.4 pg/ml. Blind duplicates were performed over the full range of ARIC participants.

Self-reported variables

Questionnaires were administered to all participants at each ARIC visit. Age, sex, and race were determined based on self-report. Education was categorized as less than a high school diploma, a high school diploma, or beyond a high school diploma; physical activity was calculated using the Baecke questionnaire for sports during leisure time (both carried forward from visit 1).18 Smoking status and alcohol intake were categorized as current, former, or never users.

Objectively measured variables

Body mass index (kg/m2) and waist-to-hip ratio (cm) were calculated from objective height (meters), weight (kg), and girth of waist and hip (cm) collected by trained ARIC study staff using standardized protocols.17 High-density lipoprotein cholesterol (HDL, mg/dL), low-density lipoprotein cholesterol (LDL, Friedewald equation, mg/dL),19 and triglycerides (mg/dL) were measured as previously reported.17 High-sensitivity C-reactive protein, parathyroid hormone, phosphorous, sodium, and calcium were measured in 2012–2013 in stored visit 2 serum (all mg/dL, Roche Diagnostics, Indianapolis, IN). Cystatin C was measured using a Roche Modular P Chemistry analyzer and Gentian cystatin C reagent; serum creatinine was determined using a modified kinetic Jaffé reaction. Kidney function (estimated glomerular filtration rate; eGFR) was derived using the CKD-EPI 2012 equation which incorporates both creatinine and cystatin C20. We modeled eGFR in two ways: continuously ml/min/1.73m2 and categorically as ≥90, ≥60 to <90, and ≥15 to <60 ml/min/1.73 m2. Prevalent coronary heart disease, heart failure, stroke, and diabetes definitions have been previously described.9,17 Finally, all medication use was ascertained by a medication inventory conducted at the baseline study visit.

Incident hypertension

Seated blood pressure was measured using a random zero sphygmomanometer. All participants rested for 5 minutes prior to blood pressure ascertainment; three measurements were collected at visit 3 (mean of final two calculated for analysis), whereas at visit 4 only two measurements were collected (mean calculated for analysis). Incident hypertension was defined as mean systolic blood pressure (SBP) of ≥ 140 mmHg, and/or diastolic blood pressure (DBP) ≥90 mmHg, and/or antihypertensive medication use reported at ARIC visits 3 or 4.21 For our secondary analysis, incident self-reported hypertension was identified through ARIC annual follow-up phone calls through 2012. At each phone call, participants were asked: (i) “Since we last contacted you, has a doctor said you had high blood pressure?” and (ii) “Did you take any medications during the past two weeks for high blood pressure?” Participants who answered affirmatively to either of these questions were classified as having incident self-reported hypertension.22

Statistical analysis

Baseline characteristics of participants are described using means and proportions stratified by categories of FGF23. For the primary analysis, complementary log-log (cloglog) regression models were used to explore the relationship between serum FGF23 and risk of incident hypertension. Cloglog models account for data that are collected in intervals but originate continuously; beta coefficients are interpreted as log hazard ratio (HR) providing ease of interpretation.23 For the secondary analysis, Cox regression models were used to examine the association between serum FGF23 and risk of self-reported incident hypertension. Restricted cubic splines using 5 knots at 5%, 25%, 50%, 75%, and 95% were used to visually depict the associations, and aid in selecting the most appropriate representation. The spline models suggested an exponential nonlinear relationship between FGF23 and incident hypertension. To elucidate the association more thoroughly, the upper FGF23 quintile was split into deciles 9 and 10.

The primary analysis included only objectively measured hypertension at study visits; a secondary analysis included self-reported hypertension obtained though study visits and annual follow-up telephone calls. Model 1 included adjustment for demographic characteristics (age, race, sex, ARIC field center, and educational attainment). Model 2 additionally adjusted for lifestyle variables (cigarette smoking, physical activity, body mass index, alcohol consumption, and serum phosphorus). Model 3 further adjusted for kidney function (eGFR). Interaction terms were used to evaluate whether race, sex, age, and kidney function (eGFR) modified the relationship between serum FGF23 and risk of incident hypertension. Sensitivity analyses explored excluding individuals with: (i) prevalent coronary heart disease (CHD), and (ii) prehypertension to evaluate the robustness of the association. Further sensitivity analyses examined the impact of adding serum parathyroid hormone, calcium, sodium, 25-hydroxyvitamin D and baseline cardiovascular medication use (cholesterol lowering medications, anticoagulants, and aspirin) to Model 3. Analyses were performed using SAS 9.3 (SAS Institute Inc., Cary, North Carolina); we used a conventional threshold of α=0.05 (a 5% Type I error rate).

RESULTS

The 7,948 participants who were free of hypertension at baseline (visit 2, 1990–1992) were on average 56.3 years old, 56% were female, and 17% African American. Mean (SD) serum FGF23 was 43.6 (±15.1) pg/mL. Serum FGF23 was positively associated with some traditional cardiovascular risk factors and prevalent cardiovascular diseases. These results were most evident in the upper deciles of FGF23. In particular, participants with high FGF23 were more likely to have elevated LDL, HDL, and triglycerides, serum calcium and phosphorus, impaired estimated glomerular filtration rate (eGFR), and prevalent diabetes (Table 1). Differences in kidney function were notable: 4.9% of participants in the top decile of FGF23 compared with 0.4% of participants in the first quintile of FGF23 had an eGFR <60 ml/min/1.73m2 (Table 1). Mean age of participants was similar across FGF23 quantiles: 56.8 ± 5.8 years in decile 10 and 55.8 ± 5.6 years in quintile 1. Likewise, no significant differences were observed between FGF23 category and race, education, physical activity (sport index), prevalent heart failure, prevalent stroke, or serum sodium.

Table 1.

Baseline (visit 2) participant characteristics by FGF23 category, ARIC, 1990–92 (n=7,948).

Serum FGF23 (pg/mL) Categories
  Category Quintile 1 Quintile 2 Quintile 3 Quintile 4 Decile 9 Decile 10 P trend1 Missing (n)
  Median 25.9 34.7 40.7 47.8 55.9 73.2
  Range 2.9 – 31.3 31.4 – 37.8 37.9 – 43.8 43.9 – 52.4 52.5 – 60.5 60.6 – 181.8
  N total 1,589 1,590 1,590 1,590 795 794
  N incident hypertension 405 424 426 416 215 266
Demographics
Age, mean years ± SD 55.8 ± 5.6 56.1 ± 5.6 56.2 ± 5.6 56.3 ± 5.6 56.5 ± 5.7 56.8 ± 5.8 0.09 0
Female n (%) 974 (61.3) 938 (59.0) 881 (55.4) 842 (53.0) 427 (53.7) 434 (54.7) <0.0001 0
African American, n (%) 301 (18.9) 261 (16.4) 271 (17.0) 256 (16.1) 130 (16.3) 125 (15.7) 0.25 0
Education, > High school 635 (40.0) 641 (40.4) 659 (41.5) 666 (42.0) 303 (38.2) 321 (40.5) 0.61 11
Behavioral Variables
Sport Index, mean ± SD 2.5 ± 0.8 2.5 ± 0.8 2.5 ± 0.8 2.5 ± 0.8 2.5 ± 0.8 2.5 ± 0.8 0.26 24
Smoking, current, n (%) 425 (26.8) 365 (23.0) 333 (21.0) 325 (20.5) 154 (19.4) 144 (18.2) <0.0001 8
Alcohol, current, n (%) 1016 (64.0) 967 (60.9) 959 (60.4) 975 (61.4) 472 (59.4) 449 (56.5) 0.04 7
Prevalent Disease
Diabetes, n (%) 134 (8.4) 124 (7.8) 140 (8.8) 145 (9.2) 87 (10.9) 91 (11.5) 0.03 18
Heart failure, n (%) 14 (0.9) 24 (1.5) 23 (1.5) 25 (1.6) 18 (2.3) 18 (2.3) 0.07 133
Stroke, n (%) 8 (0.5) 6 (0.4) 10 (0.6) 12 (0.8) 10 (1.3) 42 (5.4) 0.14 15
CHD, n (%) 47 (3.0) 49 (3.1) 73 (4.7) 61 (3.9) 37 (4.8) 8 (1.0) 0.01 150
Physiologic Variables
BMI, kg/m2 ± SD 26.1 ± 4.7 26.8 ± 4.7 27.1 ± 4.7 27.2 ± 4.6 27.6 ± 4.8 27.9 ± 5.0 <0.0001 6
LDL, mg/dL ± SD 127 ± 37 131 ± 35 133 ± 35 133 ± 35 138 ± 37 135 ± 37 <0.0001 115
HDL, mg/dL ± SD 54.0 ± 17.5 51.9 ± 16.9 50.4 ± 16.8 49.6 ± 17.0 48.4 ± 15.9 48.0 ± 16.6 <0.0001 32
Triglycerides, mg/dL ± SD 113.7 ± 73.1 119.9 ± 65.9 126.4 ± 75.6 132.1 ± 80.9 137.2 ± 75.3 148.4 ± 95.4 <0.0001 11
Phosphorous, mg/dL ±SD 3.5 ± 0.5 3.5 ± 0.5 3.5 ± 0.5 3.6 ± 0.5 3.6 ± 0.5 3.6 ± 0.5 <0.0001 53
Calcium, mg/dL ±SD 9.2 ± 0.4 9.3 ± 0.4 9.3 ± 0.4 9.3 ± 0.4 9.4 ± 0.4 9.4 ± 0.4 <0.0001 51
Sodium 140 ± 2.2 141 ± 2.2 141 ± 2.3 141 ± 2.1 141 ± 2.1 141 ± 2.2 0.68 0
eGFR, ml/min/1.73m2 ± SD 101.4 ± 14.6 99.1 ± 14.0 97.8 ± 14.3 95.9 ± 14.3 94.0 ± 15.0 90.8 ± 17.3 <0.0001 104
eGFR Creatinine level, n (%)†
 >90 ml/min/1.73m2 1239 (78.9) 1192 (75.7) 1119 (71.5) 1062 (67.8) 488 (61.9) 429 (55.1) <0.0001 0
 60–90 ml/min/1.73m2 325 (20.7) 376 (23.9) 437 (27.9) 495 (31.6) 287 (36.4) 312 (40.0) <0.0001 0
 <60 ml/min/1.73m2 6 (0.4) 7 (0.4) 9 (0.6) 10 (0.6) 13 (1.7) 38 (4.9) <0.0001 104
Open in a new tab

Abbreviations: Fibroblast growth factor 23, FGF23; Atherosclerosis Risk in Communities Study, ARIC; standard deviation, SD; coronary heart disease, CHD; body mass index, BMI; low-density lipoprotein cholesterol, LDL; high-density lipoprotein cholesterol, HDL; estimated glomerular filtration rate, eGFR.

*

Overall P test for linear trend.

†

Kidney function (estimated glomerular filtration rate; eGFR) was derived using the CKD-EPI 2012 equation which incorporates both creatinine and cystatin C.

Incident Hypertension at Study Visits

Over a median follow-up of 5.9 years, 27% of the sample (n=2,152) developed incident hypertension. A nonlinear association between serum FGF23 and incident hypertension was observed using restricted cubic splines; only persons in the highest decile of serum FGF23 had an increased risk of incident hypertension (Table 2). After adjustment for demographics (Model 1), the hazard ratio for incident hypertension was 1.27 (95% CI: 1.14, 1.42) for the highest decile of FGF23 compared to the lowest quintile. After additional adjustment for behaviors and adiposity (Model 2) the HR was 1.24 (95% CI: 1.11, 1.39). The association was further attenuated in the final model (Model 3) after adjusting for renal function 1.21 (95% CI: 1.08, 1.35) (Table 2). There was no evidence of meaningful multiplicative interaction with age, race, sex, or eGFR category.

Table 2.

Adjusted HRs and 95% CIs for objectively measured incident hypertension by FGF23 category, ARIC, 1990–98 (total n=7,948; total developing hypertension = 2,152).

Serum FGF-23 (pg/mL) Categories
Category Quintile 1 Quintile 2 Quintile 3 Quintile 4 Decile 9 Decile 10
Median 25.9 34.7 40.7 47.8 55.9 73.2
 Range 2.9 – 31.3 31.4 – 37.8 37.9 – 43.8 43.9 – 52.4 52.5 – 60.5 60.6 – 181.8
 N total 1,589 1,590 1,590 1,590 795 794
N events 405 424 426 416 215 266
Model
Model 1 1 0.97 (0.88, 1.06) 0.97 (0.89, 1.06) 0.96 (0.87, 1.05) 0.95 (0.85, 1.07) 1.27 (1.14, 1.42)
Model 2 1 0.97 (0.89, 1.07) 0.97 (0.89, 1.06) 0.95 (0.87, 1.04) 0.93 (0.83, 1.05) 1.24 (1.11, 1.39)
Model 3 1 0.98 (0.90, 1.07) 0.98 (0.89, 1.07) 0.96 (0.87, 1.05) 0.94 (0.83, 1.05) 1.21 (1.08, 1.35)
Open in a new tab

Model 1: Adjusted for age, race, sex, ARIC field center, and educational attainment.

Model 2: Adjusted for Model 1 + cigarette smoking, alcohol intake, physical activity, body mass index, and serum phosphorus.

Model 3: Adjusted for Model 2 + eGFR category.

Incident self-reported Hypertension

Of the original 7,644 participants free of hypertension at visit 2, 55% of the sample (n=4,128) developed incident hypertension over a mean of 12.3 years (through visit 5). FGF23 was positively associated with incident self-reported hypertension in a dose-response manner (Table 3). Relative to quintile 1, the demographic-adjusted (Model 1), hazard ratios for incident self-reported hypertension were 1.33 (95% CI: 1.20, 1.48) for decile 10, 1.21 (95% CI: 1.09, 1.34) for decile 9, and 1.16 (95% CI: 1.07, 1.27) for quintile 4. After additional adjustment for behaviors and adiposity (Model 2) the hazard ratios were attenuated: 1.23 (95% CI: 1.11, 1.37) for decile 10, 1.14 (95% CI: 1.03, 1.27) for decile 9, and 1.11 (95% CI: 1.02, 1.21) Final adjustment for renal function using Model 3 revealed similar results to Model 2 (Table 3).

Table 3.

Adjusted HRs and 95% CIs for self-reported incident hypertension by FGF23 category, ARIC, 1990–2012 (total n=7,644; total developing hypertension = 4,128).

Serum FGF-23 (pg/mL) Categories
Category Quintile 1 Quintile 2 Quintile 3 Quintile 4 Decile 9 Decile 10
Median 25.1 35.3 42.9 47.9 53.8 74.1
 Range 2.8 – 31.1 31.2 – 39.1 39.2 – 44.8 44.9 – 53.9 54.0 – 60.6 60.7 – 181.9
 N total 1,529 1,529 1,529 1,529 764 764
N events 803 831 828 839 399 428
Model
Model 1 1 1.10 (1.00, 1.19) 1.11 (0.98, 1.19) 1.16 (1.07, 1.27) 1.21 (1.09, 1.34) 1.33 (1.20, 1.48)
Model 2 1 1.09 (0.99, 1.16) 1.09 (0.99, 1.16) 1.11 (1.02, 1.21) 1.14 (1.03, 1.27) 1.23 (1.11, 1.37)
Model 3 1 1.09 (1.00, 1.16) 1.09 (0.99, 1.16) 1.10 (1.02, 1.20) 1.13 (1.03, 1.26) 1.21 (1.10, 1.37)
Open in a new tab

Model 1: Adjusted for age, race, sex, ARIC field center, and educational attainment.

Model 2: Adjusted for Model 1 + cigarette smoking, alcohol intake, physical activity, body mass index, and serum phosphorus.

Model 3: Adjusted for Model 2 + eGFR category.

Sensitivity Analyses

Sensitivity analyses excluding: (i) individuals with prevalent CHD (self-reported physician diagnosis of myocardial infarction (MI) or coronary revascularization (coronary artery bypass surgery or angioplasty), prevalent MI diagnosed from ECG at visit 1, or an incident adjudicated CHD event between ARIC visits 1 and 2), and (ii) individuals with pre-hypertension (systolic blood pressure of 120 to 139 mm Hg or diastolic blood pressure of 80 to 89 mm Hg) resulted in a higher risk of incident hypertension compared to the primary analysis (Table 4). When excluding individuals with CHD, individuals in the upper decile of FGF23 had a 30% higher risk of incident hypertension (HR: 1.30 95% CI: 1.11, 1.54) compared with analyses not excluding these individuals (HR: 1.21, 95% CI: 1.08, 1.35) in the fully adjusted model (Model 3). When excluding individuals with pre-hypertension, individuals in the upper decile of FGF23 had a 42% higher incidence of hypertension (HR: 1.42 95% CI: 1.06, 1.90) in the fully adjusted model. Results were nearly identical when additionally adjusting for serum parathyroid hormone, calcium, sodium, 25-hydroxyvitamin D and baseline cardiovascular medication use (in separate models, and a model with all biomarkers included; data not shown).

Table 4.

Adjusted HRs and 95% CIs for incident hypertension by FGF23 category, ARIC, sensitivity analyses for individuals without coronary heart disease (CHD)* and pre-hypertension.†

Model Quintile 1 Quintile 2 Quintile 3 Quintile 4 Decile 9 Decile 10
Main Results
Model 1 1 0.97 (0.88, 1.06) 0.97 (0.89, 1.06) 0.96 (0.87, 1.05) 0.95 (0.85, 1.07) 1.27 (1.14, 1.42)
Model 2 1 0.97 (0.89, 1.07) 0.97 (0.89, 1.06) 0.95 (0.87, 1.04) 0.93 (0.83, 1.05) 1.24 (1.11, 1.39)
Model 3 1 0.98 (0.90, 1.07) 0.98 (0.89, 1.07) 0.96 (0.87, 1.05) 0.94 (0.83, 1.05) 1.21 (1.08, 1.35)
Sensitivity Analysis: Individuals without CHD*
Model 1 1 1.08 (0.94, 1.24) 1.08 (0.94, 1.24) 1.04 (0.91, 1.20) 1.02 (0.86, 1.21) 1.43 (1.22, 1.68)
Model 2 1 1.04 (0.90, 1.19) 1.02 (0.89, 1.18) 0.99 (0.86, 1.14) 0.95 (0.80, 1.13) 1.33 (1.13, 1.57)
Model 3 1 1.04 (0.91, 1.20) 1.03 (0.89, 1.18) 1.00 (0.86, 1.15) 0.96 (0.80, 1.14) 1.30 (1.11, 1.54)
Sensitivity Analysis: Individuals without pre-hypertension†
Model 1 1 1.23 (0.96, 1.56) 1.13 (0.89, 1.44) 1.11 (0.87, 1.42) 1.54 (1.18. 2.02) 1.58 (1.20, 2.10)
Model 2 1 1.19 (0.93, 1.51) 1.08 (0.84, 1.38) 1.07 (0.83, 1.37) 1.49 (1.14, 1.95) 1.51 (1.13, 2.00)
Model 3 1 1.21 (0.95, 1.54) 1.09 (0.85, 1.40) 1.07 (0.83, 1.38) 1.49 (1.14, 1.96) 1.42 (1.06, 1.90)
Open in a new tab
*

CHD defined as self-reported physician diagnosis of myocardial infarction (MI) or coronary revascularization (coronary artery bypass surgery or angioplasty), prevalent MI diagnosed from ECG at visit 1, or an incident adjudicated CHD event between ARIC visits 1 and 2.

†

Pre-hypertension defined as systolic blood pressure 120–139 mm Hg or diastolic blood pressure of 80–89 mm Hg.

Model 1: Adjusted for age, race, sex, ARIC field center, and educational attainment.

Model 2: Adjusted for Model 1 + cigarette smoking, alcohol intake, physical activity, body mass index, and serum phosphorus.

Model 3: Adjusted for Model 2 + eGFR category.

DISCUSSION

In this large community-based cohort, a non-linear association was found between FGF23 and risk of incident hypertension, independent of traditional cardiovascular risk factors and indicators of kidney function. Participants in the highest decile of FGF23 were at approximately 21% greater risk of incident hypertension after adjustment for demographics, health behaviors, and traditional cardiovascular risk factors. These associations were independent of kidney function. Furthermore, the results were similar regardless of whether incident hypertension was defined through objective measurements or by self-report, with the highest decile of FGF23 being associated with an elevated risk of incident hypertension for both methods of measurement. These findings provide human evidence in a large cohort study that FGF23 is associated with the development of incident hypertension, thereby extending the current understanding of the role of FGF23 on cardiovascular risk. Our findings are supported by a recent study which reported that FGF23 was correlated with prevalent hypertension in children and adolescents.24

Recent evidence from human and animal studies provides insight into the mechanisms through which elevated FGF23 may increase the risk of incident hypertension. Specifically, FGF23 has been found to initiate mineral homeostasis disruption, and a consequent increase in blood pressure. In mouse models, FGF23 was found to increase calcium and sodium reabsorption, as well as renal phosphate excretion.13,14 Though the literature are complex, it is hypothesized that calcium reabsorption may lead to serum calcium accumulation, vascular calcification, and consequent elevated blood pressure.25,26 A recent cross-sectional analysis of National Health and Nutrition Examination Survey data demonstrated a positive association between serum calcium and prevalent hypertension; individuals in the highest quartile of serum calcium had a 1.49 higher odds of hypertension (95% CI: 1.15–1.93, p=0.005) than those in the lowest quartile.27 The effects of sodium on blood volume have been extensively documented independent of FGF23.28–33 In a randomized controlled trial using the Dietary Approaches to Stop Hypertension (DASH) diet, blood pressure was reduced by 11.5 and 7.1 mm Hg in participants with and without hypertension respectively.34 Researchers recently found that FGF23 is an influential regulator of sodium reabsorption in the kidney and overall plasma volume, an important contributor to blood pressure. Mice that were lacking the FGF23 gene excreted more sodium in their urine, whereas mice with elevated FGF23 levels had higher plasma sodium and blood pressure.13 Furthermore, though a rise in serum phosphorus precedes FGF23 elevation,35 data from epidemiological studies are equivocal in establishing the relationship between serum phosphate levels and blood pressure.36 In our analyses, the impact of adjusting for serum phosphorus, sodium, and calcium (in separate models, and a model with all biomarkers included) was negligible. To further complicate matters, markers of mineral metabolism are highly correlated; there is a paucity of data available on the combined effects of phosphorus, calcium, and sodium on blood pressure. As a result, the ability to specify the impact of one mineral on blood pressure in isolation is limited.37

Additional mechanisms exist linking FGF23 and hypertension independent of mineral homeostasis. It is possible that FGF23 is an early indicator of impaired kidney function, an established risk factor for incident hypertension.38 Although the current analyses adjusted for kidney function, FGF23 may be elevated prior to frank changes in the kidney. In addition, a positive association has been found between FGF23 and atherosclerosis,39 though it is unclear whether it acts independently of phosphorus.25,40 Finally, high circulating FGF23 has been associated with endothelial dysfunction and arterial wall stiffness,41–43 and inflammation44,45 - factors which may contribute to the increased peripheral resistance typical of essential hypertension.

Strengths of this study include the prospective design with ample long-term follow-up, a community-based bi-racial cohort, a large sample size, and a comprehensive collection of cardiovascular risk factors for confounding adjustment. Importantly, biologically active and intact FGF23 and incident hypertension were both objectively measured for our primary analyses. Secondary analyses using self-reported hypertension with much longer follow-up (mean 12.3 vs. 5.9 years) yielded similar results. Limitations of this study include that FGF23 was measured at a single point in time. Any associated measurement error, however, was likely non-differential therefore attenuating the results of the analyses towards the null. Second, Klotho, a co-receptor for FGF23, was not measured in the present study. Klotho is necessary for FGF23 to have biological activity;5 elevated FGF23 levels are considered more informative when presence of Klotho is verified. Third, the possibility of selection bias exists. It is possible that those who participated in the present study differ in their association of FGF23 and hypertension compared to those lost to follow-up. Finally, despite our attempts to statistically adjust for known confounders, residual confounding may be present.

Identification of new biomarkers that contribute to hypertension provides the opportunity to identify individuals at high risk, develop pharmacologic agents for intervention, and improve our understanding of hypertension physiology. Clinical identification of patients who are high risk for hypertension could aid in earlier identification, risk classification, and more targeted treatment. Furthermore, pharmacologic interventions to reduce the production or biological activity of FGF23 – which are primarily being created with the goal of preventing both cardiovascular and kidney disease – are in various stages of development. Oral phosphate binders, FGF23 monoclonal antibodies, and FGF receptor agonists are treatment options being considered, particularly in the chronic kidney disease population.25,36 Oral phosphate binders are the only therapeutic agent currently available, and have been shown to effectively reduce FGF23 levels after 6 weeks of treatment.25,35,46–49 Additional investigations with longer follow-up times are needed. Clinical trials with hard endpoints are of utmost importance to determine whether these agents hold therapeutic value.

In conclusion, in this large community-based cohort elevated FGF23 levels were associated with a greater risk of incident hypertension, independent of traditional hypertension and cardiovascular risk factors and abnormal kidney function. These data suggest FGF23 as a possible risk marker for incident hypertension. Next steps include replication in other cohorts and further exploration of the biological mechanisms underlying this association.

Acknowledgments

Sources of support:

Measurement of FGF23 and some of the related biomarkers was supported through a grant from the National Heart, Lung, and Blood Institute, R01 HL103706 (Lutsey, PI). Cystatin C and CRP were measured as part of a grant from the National Institute of Diabetes and Digestive and Kidney Disease R01 DK089174 (Selvin, PI). The Atherosclerosis Risk in Communities Study is a multi-center collaborative study supported by National Heart, Lung, and Blood Institute contracts, HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C. The trainee was supported by a National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Number T32HL07779 (Folsom, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The authors thank the staff and participants of the ARIC study for their important contributions. Roche Diagnostics donated the reagents for the C-reactive protein assays used in this study.

Footnotes

Previous presentations: AHA EPI/Lifestyle 2014, poster presentation

Conflicts of interest: None

References

  • 1.Li C, Balluz LS, Ford ES, Okoro CA, Zhao G, Pierannunzi C. A comparison of prevalence estimates for selected health indicators and chronic diseases or conditions from the Behavioral Risk Factor Surveillance System, the National Health Interview Survey, and the National Health and Nutrition Examination Survey, 200. Prev Med (Baltim) 2012;54(6):381–7. doi: 10.1016/j.ypmed.2012.04.003. [DOI] [PubMed] [Google Scholar]
  • 2.Crescioni M, Gorina Y, Bilheimer L, Gillum RF. Trends in health status and health care use among older men. Natl Health Stat Report. 2010;(24):1–18. [PubMed] [Google Scholar]
  • 3.Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics--2013 update: a report from the American Heart Association. Circulation. 2013;127(1):e6–e245. doi: 10.1161/CIR.0b013e31828124ad. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Lifton RP, Gharavi AG, Geller DS. Molecular mechanisms of human hypertension. Cell. 2001;104(4):545–56. doi: 10.1016/s0092-8674(01)00241-0. [DOI] [PubMed] [Google Scholar]
  • 5.Hu MC, Shiizaki K, Kuro-o M, Moe OW. Fibroblast growth factor 23 and Klotho: physiology and pathophysiology of an endocrine network of mineral metabolism. Annu Rev Physiol. 2013;75:503–33. doi: 10.1146/annurev-physiol-030212-183727. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Faul C. Fibroblast growth factor 23 and the heart. Curr Opin Nephrol Hypertens. 2012;21(4):369–75. doi: 10.1097/MNH.0b013e32835422c4. [DOI] [PubMed] [Google Scholar]
  • 7.Dalal M, Sun K, Cappola AR, et al. Relationship of serum fibroblast growth factor 23 with cardiovascular disease in older community-dwelling women. Eur J Endocrinol. 2011;165(5):797–803. doi: 10.1530/EJE-11-0577. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ärnlöv J, Carlsson AC, Sundström J, et al. Serum FGF23 and risk of cardiovascular events in relation to mineral metabolism and cardiovascular pathology. Clin J Am Soc Nephrol. 2013;8(5):781–6. doi: 10.2215/CJN.09570912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Lutsey PL, Alonso A, Selvin E, et al. Fibroblast growth factor-23 and incident coronary heart disease, heart failure, and cardiovascular mortality: the atherosclerosis risk in communities study. J Am Heart Assoc. 2014;3(3):e000936. doi: 10.1161/JAHA.114.000936. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Isakova T, Xie H, Yang W, et al. Fibroblast growth factor 23 and risks of mortality and end-stage renal disease in patients with chronic kidney disease. JAMA. 2011;305(23):2432–9. doi: 10.1001/jama.2011.826. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Gutiérrez OM, Mannstadt M, Isakova T, et al. Fibroblast growth factor 23 and mortality among patients undergoing hemodialysis. N Engl J Med. 2008;359(6):584–92. doi: 10.1056/NEJMoa0706130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Kendrick J, Cheung AK, Kaufman JS, et al. FGF-23 associates with death, cardiovascular events, and initiation of chronic dialysis. J Am Soc Nephrol. 2011;22(10):1913–22. doi: 10.1681/ASN.2010121224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Andrukhova O, Slavic S, Smorodchenko A, et al. FGF23 regulates renal sodium handling and blood pressure. EMBO Mol Med. 2014;6(6):744–59. doi: 10.1002/emmm.201303716. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Andrukhova O, Smorodchenko A, Egerbacher M, et al. FGF23 promotes renal calcium reabsorption through the TRPV5 channel. EMBO J. 2014;33(3):229–46. doi: 10.1002/embj.201284188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Parker BD, Schurgers LJ, Brandenburg VM, et al. The associations of fibroblast growth factor 23 and uncarboxylated matrix Gla protein with mortality in coronary artery disease: the Heart and Soul Study. Ann Intern Med. 2010;152(10):640–8. doi: 10.1059/0003-4819-152-10-201005180-00004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ix JH, Katz R, Kestenbaum BR, et al. Fibroblast growth factor-23 and death, heart failure, and cardiovascular events in community-living individuals: CHS (Cardiovascular Health Study) J Am Coll Cardiol. 2012;60(3):200–7. doi: 10.1016/j.jacc.2012.03.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators. Am J Epidemiol. 1989;129(4):687–702. [PubMed] [Google Scholar]
  • 18.Baecke JA, Burema J, Frijters JE. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36(5):936–42. doi: 10.1093/ajcn/36.5.936. [DOI] [PubMed] [Google Scholar]
  • 19.Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18(6):499–502. [PubMed] [Google Scholar]
  • 20.Inker LA, Schmid CH, Tighiouart H, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367(1):20–9. doi: 10.1056/NEJMoa1114248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Atherosclerosis Risk in Communities (ARIC) Study Operations Manual 11, Visit 4. 1997.
  • 22.Bower JK, Appel LJ, Matsushita K, et al. Glycated hemoglobin and risk of hypertension in the atherosclerosis risk in communities study. Diabetes Care. 2012;35(5):1031–7. doi: 10.2337/dc11-2248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Allison PD. Survival Analysis Using SAS: A Practical Guide. 12. Cary, NC: SAS Publishing; 2008. [Google Scholar]
  • 24.Zając M, Rybi-Szumińska A, Wasilewska A. Urine fibroblast growth factor 23 levels in hypertensive children and adolescents. Croat Med J. 2015;56(4):344–50. doi: 10.3325/cmj.2015.56.344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Heine GH, Seiler S, Fliser D. FGF-23: the rise of a novel cardiovascular risk marker in CKD. Nephrol Dial Transplant. 2012;27(8):3072–81. doi: 10.1093/ndt/gfs259. [DOI] [PubMed] [Google Scholar]
  • 26.Foley RN, Collins AJ, Herzog CA, Ishani A, Kalra PA. Serum phosphorus levels associate with coronary atherosclerosis in young adults. J Am Soc Nephrol. 2009;20(2):397–404. doi: 10.1681/ASN.2008020141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Sabanayagam C, Shankar A. Serum calcium levels and hypertension among U.S. adults. J Clin Hypertens (Greenwich) 2011;13(10):716–21. doi: 10.1111/j.1751-7176.2011.00503.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Intersalt: an international study of electrolyte excretion and blood pressure. Results for 24 hour urinary sodium and potassium excretion. Intersalt Cooperative Research Group. BMJ. 1988;297(6644):319–28. doi: 10.1136/bmj.297.6644.319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Elliott P, Stamler J, Nichols R, et al. Intersalt revisited: further analyses of 24 hour sodium excretion and blood pressure within and across populations. Intersalt Cooperative Research Group. BMJ. 1996;312(7041):1249–53. doi: 10.1136/bmj.312.7041.1249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Graudal NA, Galløe AM, Garred P. Effects of sodium restriction on blood pressure, renin, aldosterone, catecholamines, cholesterols, and triglyceride: a meta-analysis. JAMA. 1998;279(17):1383–91. doi: 10.1001/jama.279.17.1383. [DOI] [PubMed] [Google Scholar]
  • 31.Midgley JP, Matthew AG, Greenwood CM, Logan AG. Effect of reduced dietary sodium on blood pressure: a meta-analysis of randomized controlled trials. JAMA. 275(20):1590–7. doi: 10.1001/jama.1996.03530440070039. [DOI] [PubMed] [Google Scholar]
  • 32.Appel LJ, Moore TJ, Obarzanek E, et al. A clinical trial of the effects of dietary patterns on blood pressure. DASH Collaborative Research Group. N Engl J Med. 1997;336(16):1117–24. doi: 10.1056/NEJM199704173361601. [DOI] [PubMed] [Google Scholar]
  • 33.Whelton PK, He J, Cutler JA, et al. Effects of oral potassium on blood pressure. Meta-analysis of randomized controlled clinical trials. JAMA. 1997;277(20):1624–32. doi: 10.1001/jama.1997.03540440058033. [DOI] [PubMed] [Google Scholar]
  • 34.Sacks FM, Svetkey LP, Vollmer WM, et al. 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(1):3–10. doi: 10.1056/NEJM200101043440101. [DOI] [PubMed] [Google Scholar]
  • 35.Bergwitz C, Jüppner H. Regulation of phosphate homeostasis by PTH, vitamin D, and FGF23. Annu Rev Med. 2010;61:91–104. doi: 10.1146/annurev.med.051308.111339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Jüppner H, Wolf M, Salusky IB. FGF-23: More than a regulator of renal phosphate handling? J Bone Miner Res. 2010;25(10):2091–7. doi: 10.1002/jbmr.170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Elliott P, Kesteloot H, Appel LJ, et al. Dietary phosphorus and blood pressure: international study of macro- and micro-nutrients and blood pressure. Hypertension. 2008;51(3):669–75. doi: 10.1161/HYPERTENSIONAHA.107.103747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Salem MM. Pathophysiology of hypertension in renal failure. Semin Nephrol. 2002;22(1):17–26. [PubMed] [Google Scholar]
  • 39.Zoccali C, Yilmaz MI, Mallamaci F. FGF23: A Mature Renal and Cardiovascular Risk Factor? Blood Purif. 2013;36(1):52–7. doi: 10.1159/000351001. [DOI] [PubMed] [Google Scholar]
  • 40.Wolf M. Update on fibroblast growth factor 23 in chronic kidney disease. Kidney Int. 2012;82(7):737–47. doi: 10.1038/ki.2012.176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Stevens KK, McQuarrie EP, Sands W, et al. Fibroblast growth factor 23 predicts left ventricular mass and induces cell adhesion molecule formation. Int J Nephrol. 2011;2011:297070. doi: 10.4061/2011/297070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Mirza MAI, Larsson A, Lind L, Larsson TE. Circulating fibroblast growth factor–23 is associated with vascular dysfunction in the community. Atherosclerosis. 2009;205(2):385–90. doi: 10.1016/j.atherosclerosis.2009.01.001. [DOI] [PubMed] [Google Scholar]
  • 43.Yilmaz MI, Sonmez A, Saglam M, et al. FGF-23 and vascular dysfunction in patients with stage 3 and 4 chronic kidney disease. Kidney Int. 2010;78(7):679–85. doi: 10.1038/ki.2010.194. [DOI] [PubMed] [Google Scholar]
  • 44.Manghat P, Fraser WD, Wierzbicki AS, Fogelman I, Goldsmith DJ, Hampson G. Fibroblast growth factor-23 is associated with C-reactive protein, serum phosphate and bone mineral density in chronic kidney disease. Osteoporos Int. 2010;21(11):1853–61. doi: 10.1007/s00198-009-1142-4. [DOI] [PubMed] [Google Scholar]
  • 45.Munoz Mendoza J, Isakova T, Ricardo AC, et al. Fibroblast growth factor 23 and Inflammation in CKD. Clin J Am Soc Nephrol. 2012;7(7):1155–62. doi: 10.2215/CJN.13281211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Koiwa F, Kazama JJ, Tokumoto A, et al. Sevelamer hydrochloride and calcium bicarbonate reduce serum fibroblast growth factor 23 levels in dialysis patients. Ther Apher Dial. 2005;9(4):336–9. doi: 10.1111/j.1744-9987.2005.00293.x. [DOI] [PubMed] [Google Scholar]
  • 47.Nagano N, Miyata S, Abe M, et al. Effect of manipulating serum phosphorus with phosphate binder on circulating PTH and FGF23 in renal failure rats. Kidney Int. 2006;69(3):531–7. doi: 10.1038/sj.ki.5000020. [DOI] [PubMed] [Google Scholar]
  • 48.Pande S, Ritter CS, Rothstein M, et al. FGF-23 and sFRP-4 in chronic kidney disease and post-renal transplantation. Nephron Physiol. 2006;104(1):p23–32. doi: 10.1159/000093277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Isakova T, Gutiérrez OM, Smith K, et al. Pilot study of dietary phosphorus restriction and phosphorus binders to target fibroblast growth factor 23 in patients with chronic kidney disease. Nephrol Dial Transplant. 2011;26(2):584–91. doi: 10.1093/ndt/gfq419. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES