Abstract
It is unclear whether metabolic syndrome (MetS) is associated with atrial fibrillation (AF) in an older population with greater cardiovascular risk, including those with chronic kidney disease. The authors investigated the association between MetS and AF in participants in SPRINT (Systolic Blood Pressure Intervention Trial). MetS was defined based on the Modified Third National Cholesterol Education Program. The baseline prevalence rate for MetS was 55%, while 8.2% of the participants had AF. In multivariate regression analyses, AF was not associated with presence of MetS in either chronic kidney disease or non–chronic kidney disease subgroups. Age, race, history of cardiovascular diseases, decreased triglycerides, decreased pulse pressure, and albuminuria remained significantly associated with AF risk. In contrast to the general population, MetS was not associated with AF in the older population with increased cardiovascular risk studied in SPRINT.
Keywords: chronic renal failure, insulin resistance, metabolic syndrome
1. INTRODUCTION
Atrial fibrillation (AF) is a significant public health problem for both the general population and patients with chronic kidney disease (CKD). It is associated with a 66% increase in the relative rate of death in patients with CKD not on dialysis.1 Incident AF in patients with CKD is also associated with a 67% increase in the rate of end‐stage renal disease.2 The known independent predictors for AF in the CKD population, including age, white race, increased left atrial diameter, and congestive heart failure,3 are not comprehensive and do not present readily treatable targets, underscoring a critical need for novel approaches to risk stratification and better understanding of the pathomechanism.
AF is increasingly thought to be a metabolic cardiac disease, with much of the recent data establishing a significant association between insulin resistance and AF risk in the general population. Individual components of metabolic syndrome (MetS), which are associated with insulin resistance, have all been associated with increased risk for AF.4, 5 Diabetes mellitus in particular has been shown to be independently associated with an increased risk of AF,6 with significant correlation with the duration of diabetes mellitus and the level of dysglycemia.7, 8 In particular, increased fasting insulin levels independently predict AF, with each increment of 18 mg/dL being associated with a 33% increased risk of incident AF.9 Despite numerous studies suggesting an important role of glucose and insulin homeostasis in AF risk in the general population, the underlying mechanisms are incompletely understood. In the general population, impaired fasting glucose is associated with an intra‐atrial conduction delay and greater AF recurrence rate after catheter ablation.10 In addition, fasting glucose level is independently associated with impaired endothelial function and increased arterial stiffness, factors associated with increased risk of AF.11 Hyperinsulinemia also causes activation of the renin‐angiotensin‐aldosterone and sympathetic systems, which are associated with increased risk for cardiac fibrosis and arrhythymia.12
Thus, dysglycemia appears to play an important role in the pathogenesis of AF. The specific role of impaired fasting glucose in AF, however, has not been investigated in the older population with additional cardiovascular risk, including patients with CKD. The AF prevalence rate is increased two‐ to three‐fold in CKD.13 Compared with the general population in which the prevalence rates for MetS and impaired fasting glucose are about 20%,14 the corresponding prevalence is also increased two‐ to three‐fold in nondiabetic patients with CKD.15, 16 Furthermore, insulin resistance and glucose intolerance are significantly associated with left ventricular hypertrophy, a feature frequently associated with AF risk, in CKD stages 1 through 3,17 and cardiovascular mortality in end‐stage renal disease.18 It is plausible that the significantly increased AF risk in CKD cohorts can be partially explained by the similarly increased prevalence of dysglycemia. Recently available data consistently demonstrate that angiotensin‐converting enzyme inhibitors, which have insulin‐sensitizing effect, can significantly reduce AF incidence.19, 20 Given the exceedingly high burden of cardiovascular mortality in the CKD population, particularly in association with AF, it is important to develop a better understanding of its pathophysiology. We hypothesize that MetS is associated with increased AF risk in participants in SPRINT (Systolic Blood Pressure Intervention Trial), a prospective, randomized clinical trial targeting older individuals with hypertension and increased cardiovascular risk, including CKD. Using the cross‐sectional analyses of the baseline characteristics of the SPRINT participants, we sought to characterize the association between metabolic profile and AF risk in CKD and non‐CKD SPRINT subcohorts.
2. METHODS
2.1. Study population
SPRINT is a multicenter, randomized, controlled trial to test the primary hypothesis that treating to a systolic blood pressure (BP) target of <120 mm Hg (intensive intervention) compared with a systolic BP target of <140 mm Hg (standard intervention) would reduce the composite outcome of fatal and nonfatal cardiovascular events in a nondiabetic hypertensive cohort with one other risk factor for cardiovascular disease (CVD). The SPRINT cohort thus represents a relatively high‐risk population, with recruitment targeting for three subgroups: (1) those with CKD, defined as estimated glomerular filtration rate [eGFR] 20–59 mL/min/1.73 m2); (2) those with clinical or subclinical CVD or increased cardiovascular risk, including 10‐year Framingham risk ≥15%; and (3) seniors (age ≥75 years). SPRINT met an additional target of enrolling 40% minorities (blacks 29.9%, Hispanics 10.5%, other 1.9%) but enrolled fewer women (35.6%) than the goal of 50%. The details regarding the study design and methods have been previously described.21, 22
For the purposes of our study, we limited our analyses to the SPRINT participants with all available covariates required for evaluation of MetS and kidney function, including fasting serum glucose levels, serum lipid levels, and urine albumin‐creatinine ratio. Of the entire 9361 original SPRINT patients, 1098 participants had missing values for one or more covariates. Our study population is thus limited to 8263 SPRINT participants.
2.2. Definitions
The fasting insulin levels, necessary for the calculation of homeostatic model assessment to estimate insulin sensitivity and β‐cell function,23 were not available in SPRINT. We have thus used the following standard definitions for MetS and CKD:
-
1
MetS was defined based on the Modified Third National Cholesterol Education Program Expert Panel criteria as having ≥3 of the following: (1) body mass index (BMI) >30 kg/m2, (2) serum high‐density lipoprotein (HDL) cholesterol <50 mg/dL for women and <40 mg/dL for men, (3) serum total triglycerides >150 mg/dL, (4) systolic BP ≥130 mm Hg or diastolic BP ≥85 mm Hg or on treatment for hypertension, and (5) fasting serum glucose ≥100 mg/dL.24 We used BMI >30 kg/m2 as a substitute for waist circumference measure, which was not collected in SPRINT.
-
2
CKD was defined as eGFR, using the Modification of Diet in Renal Disease equation, <60 mL/min/1.73 m2
-
3
AF was defined by its presence on baseline ECG or medical history of AF. All randomized participants were asked whether they had been told by a physician that they have AF or flutter at baseline.
2.3. Statistical analyses
Baseline characteristics were described using means and SDs of continuous factors, and frequency and percentages of categorical factors. Univariable tests of relationships between baseline factors and prevalent AF were performed using t tests and chi‐square tests. Univariable and multivariable logistic regression models with AF as the dependent variable were used to examine the strength of association between risk factors and AF. The models included a variety of clinically relevant covariates as the predictor variables in non‐CKD and CKD SPRINT participants. These models were adjusted for the following: age, sex, race, BMI, smoking status, alcohol consumption, diabetes mellitus status, coronary heart disease, heart failure, stroke, BP, serum lipid levels (total cholesterol, triglycerides, and HDL and low‐density lipoprotein [LDL] cholesterol), eGFR, albuminuria, fasting glucose levels, and medication use (statin, niacin, fibrates, angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers). Covariates with highly skewed distribution, including urine albumin‐creatinine ratio, serum triglycerides, and serum triglycerides‐HDL ratio, were normalized using the natural‐log transformation.
3. RESULTS
3.1. The association between AF risk and MetS
The baseline demographic and clinical characteristics of the study population are described in Table 1. Of 8263 participants, 677 (8.2%) had AF at baseline. The AF diagnosis was mostly made by clinical history (predominantly by self‐report, 81.6%), and fewer AF cases were identified by the combination of clinical history and ECG criteria (15.7%) and ECG criteria alone (2.7%). The prevalence of AF was equally distributed among tertiles of the baseline systolic BP. Participants with AF were more likely to be men, older, white, and smokers. They were also more likely to have a history of CVDs, lower eGFR, increased urinary albumin to creatinine ratio, and increased pulse pressure (PP). Participants with the highest tertile of PP had the greatest prevalence of AF (9.6%), compared with a prevalence rate of 6.9% in participants with the lowest tertile of PP.
Table 1.
Distribution of selected covariates and univariable tests of association with atrial fibrillation in SPRINT participants at randomization
| Factor | Overall (N=8263) | Atrial fibrillationa (n=677) | No atrial fibrillation (n=7586) | P value |
|---|---|---|---|---|
| Demographic characteristics | ||||
| Age, y | 67.9±9.4 | 72.3±8.9 | 67.5±9.4 | <.0001 |
| Women | 2920 (35.3) | 186 (27.5) | 2734 (36.0) | <.0001 |
| Race | ||||
| Black | 2483 (30.1) | 109 (16.1) | 2374 (31.3) | <.0001 |
| Hispanic | 894 (10.8) | 33 (4.9) | 861 (11.4) | |
| Other | 144 (1.7) | 15 (2.2) | 129 (1.7) | |
| White | 4742 (57.4) | 520 (76.8) | 4222 (55.7) | |
| Clinical characteristics | ||||
| Metabolic syndrome | 4530 (54.8) | 393 (58.1) | 4137 (54.5) | .078 |
| Diabetes mellitus (by history of FBG >126) | 389 (4.7) | 38 (5.6) | 351 (4.6) | .2 |
| Coronary artery disease | 1507 (18.2) | 262 (38.7) | 1245 (16.4) | <.0001 |
| Congestive heart failure | 294 (3.6) | 94 (13.9) | 200 (2.6) | <.0001 |
| Stroke/TIA | 275 (3.3) | 45 (6.6) | 230 (3.0) | <.0001 |
| Body mass index, kg/m2 | 29.9±5.8 | 29.5±5.4 | 29.9±5.8 | .086 |
| Current or former smoker | 4838 (56.1) | 438 (61.6) | 4424 (55.5) | .002 |
| Systolic BP, mm Hg | 139.7±15.5 | 139.1±16.3 | 139.7±15.5 | .4 |
| Diastolic BP, mm Hg | 78.1±11.9 | 75.5±12.2 | 78.3±11.9 | <.0001 |
| Pulse pressure, mm Hg | 61.6±14.4 | 63.6±15.5 | 61.4±14.3 | .0004 |
| eGFR, mL/min/1.72 m2 | 71.8±20.6 | 67.1±20.1 | 72.2±20.6 | <.0001 |
| K/DOQI CKD stage | ||||
| None (eGFR ≥90 and UACR <30) | 1233 (14.9) | 63 (9.3) | 1170 (15.4) | <.0001 |
| Stage 1 (eGFR ≥90 with UACR ≥30) | 202 (2.4) | 14 (2.1) | 188 (2.5) | |
| Stage 2 (eGFR 60–89) | 4473 (54.1) | 340 (50.2) | 4133 (54.5) | |
| Stage 3A (eGFR 45–59) | 1569 (19.0) | 165 (24.4) | 1404 (18.5) | |
| Stage 3B (eGFR 30–44) | 651 (7.9) | 87 (12.9) | 564 (7.4) | |
| Stage 4 (eGFR 15–29)b | 135 (1.6) | 8 (1.7) | 127 (1.7) | |
| Fasting glucose, mg/dL | 98.9±13.7 | 100.0±13.9 | 98.8±13.7 | .037 |
| Total cholesterol, mg/dL | 189.5±40.6 | 178.0±39.7 | 190.5±40.5 | <.0001 |
| LDL cholesterol, mg/dL | 112.3±35.1 | 102.4±32.9 | 113.2±35.1 | <.0001 |
| HDL cholesterol, mg/dL | 53.0±14.4 | 52.9±14.3 | 53.0±14.7 | .9 |
| Triglycerides, mg/dL | 121.3±61.7 | 113.6±59.0 | 122.0±61.9 | .0007 |
| Triglyceride to HDL ratio | 2.59±1.79 | 2.43±1.74 | 2.60±1.80 | .014 |
| Loge (triglyceride to HDL ratio) | 0.75±0.64 | 0.69±0.62 | 0.75±0.64 | .009 |
| UACr, mg/g | 42.4±168.2 | 61.9±224.6 | 40.7±162.1 | .016 |
| Loge (UACr) | 2.56±1.18 | 2.87±1.30 | 2.53±1.16 | <.0001 |
| Albuminuria | ||||
| None | 6669 (80.7) | 483 (71.3) | 6186 (81.5) | <.0001 |
| Microalbuminuria (UACr 30–300 mg/g) | 1365 (16.5) | 163 (24.1) | 1202 (15.8) | |
| Macroalbuminuria (UACr >300 mg/g) | 229 (2.8) | 31 (4.6) | 198 (2.6) | |
| On statin | 3637 (44.0) | 387 (57.2) | 3250 (42.8) | <.0001 |
| On niacin or fenofibrate | 195 (2.4) | 22 (3.2) | 173 (2.3) | .11 |
| On antithrombotic agent (eg, aspirin and plavix) | 771 (9.3) | 277 (40.9) | 494 (6.5) | <.0001 |
| On anticoagulant (eg, warfarin and heparin) | 312 (3.8) | 212 (31.3) | 100 (1.3) | <.0001 |
| ACEI/ARB therapy | 4771 (57.7) | 426 (62.9) | 4345 (57.3) | .004 |
| Moderate or heavy drinker (>155 drinks/y) | 2616 (31.7) | 242 (35.7) | 2374 (31.3) | .017 |
Values are expressed as mean±SD for continuous factors and frequency (percentage) for categorical factors.
Atrial fibrillation diagnosis by medical history or baseline study ECG (81.6% by self‐report alone, 15.7% by combination of ECG and self‐report, and 2.7% with ECG alone).
Includes one participant with stage 5 chronic kidney disease (CKD; eGFR of 14.7 mL/min/1.73 m2). Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; BP, blood pressure; FBG, fasting blood glucose; LDL, low‐density lipoprotein; SPRINT, Systolic Blood Pressure Intervention Trial; TIA, transient ischemic attack; UACr, urine albumin to creatinine ratio.
The baseline prevalence of MetS was 55% (n=4530) in the SPRINT cohort. The clinical characteristics of the SPRINT participants with and without MetS are summarized in Table 2. Participants with MetS were younger with higher BMI and greater prevalence of coronary artery disease and heart failure. MetS was also associated with lower systolic BP and PP, higher fasting serum glucose levels, and greater albuminuria. The SPRINT participants with MetS were more likely to be treated with lipid‐lowering agents, antithrombotic agents, and angiotensin‐converting enzyme inhibitor/angiotensin receptor blocker therapy. While the prevalence of MetS was not significantly increased in participants with AF compared with those without AF (58.1% vs 54.5%, respectively; P=.078), patients with prediabetes (fasting glucose levels 100–125 mg/dL) were associated with greater prevalence of AF (9.1%) compared with those with normal fasting glucose levels (7.6%).
Table 2.
Distribution of selected covariates and univariable tests of association with metabolic syndrome in SPRINT participants at randomization
| Factor | MetS (n=4530) | No MetS (n=3736) | P value |
|---|---|---|---|
| Demographic characteristics | |||
| Age, y | 67.2±9.1 | 68.9±9.7 | <.0001 |
| Women | 1510 (33.3) | 3736 (37.7) | <.0001 |
| Race | |||
| Black | 1235 (27.3) | 1249 (33.4) | <.0001 |
| Hispanic | 505 (11.2) | 389 (10.4) | |
| Other | 75 (1.7) | 69 (1.9) | |
| White | 2715 (59.9) | 2029 (54.3) | |
| Clinical characteristics | |||
| Diabetes mellitus (by history of FBG >126) | 348 (7.7) | 41 (1.1) | <.0001 |
| Coronary artery disease | 1005 (22.2) | 502 (13.4) | <.0001 |
| Congestive heart failure | 188 (4.2) | 106 (2.8) | .0013 |
| Stroke/TIA | 141 (3.1) | 135 (3.6) | .2 |
| Body mass index, kg/m2 | 32.2±5.7 | 27.1±4.6 | <.0001 |
| Current or former smoker | 2569 (56.7) | 2072 (55.5) | .3 |
| Systolic BP, mm Hg | 138.5±15.3 | 141.1±15.8 | <.0001 |
| Diastolic BP, mm Hg | 78.0±11.9 | 78.2±12.0 | .4 |
| Pulse pressure, mm Hg | 60.5±14.1 | 62.9±14.7 | <.0001 |
| eGFR, mL/min/1.72 m2 | 71.4±20.3 | 72.2±20.9 | .058 |
| K/DOQI CKD stage | |||
| None (eGFR ≥90 and UACR <30) | 627 (13.8) | 606 (16.2) | .015 |
| Stage 1 (eGFR ≥90 with UACR ≥30) | 123 (2.7) | 79 (2.1) | |
| Stage 2 (eGFR 60–89) | 2449 (54.1) | 2028 (54.3) | |
| Stage 3A (eGFR 45–59) | 892 (19.7) | 676 (18.1) | |
| Stage 3B (eGFR 30–44) | 360 (8.0) | 291 (7.8) | |
| Stage 4 (eGFR 15–29)a | 79 (1.7) | 56 (1.5) | |
| Fasting glucose, mg/dL | 104.2±14.7 | 92.5±8.9 | <.0001 |
| Total cholesterol, mg/dL | 185.1±41.4 | 194.9±38.9 | <.0001 |
| LDL cholesterol, mg/dL | 109.7±35.6 | 116.2±34.0 | <.0001 |
| HDL cholesterol, mg/dL | 48.0±12.1 | 59.0±14.8 | <.0001 |
| Triglycerides, mg/dL | 140.3±67.9 | 98.3±43.3 | <.0001 |
| Triglyceride to HDL ratio | 3.23±2.02 | 1.81±1.02 | <.0001 |
| UACr, mg/g | 46.4±190.0 | 37.6±136.9 | .015 |
| Loge (UACr) | 2.58±1.21 | 2.53±1.14 | .062 |
| Albuminuria | |||
| None | 3609 (79.7) | 3061 (81.9) | .032 |
| Microalbuminuria (UACr 30–300 mg/g) | 786 (17.4) | 581 (15.6) | |
| Macroalbuminuria (UACr >300 mg/g) | 135 (3.0) | 94 (2.5) | |
| On statin | 2574 (56.8) | 1064 (28.5) | <.0001 |
| On niacin or fenofibrate | 150 (3.3) | 46 (1.2) | <.0001 |
| On antithrombotic agent (eg, aspirin and plavix) | 480 (10.6) | 292 (7.8) | <.0001 |
| On anticoagulant (eg, warfarin and heparin) | 195 (4.3) | 118 (3.2) | .007 |
| ACEI/ARB therapy | 2744 (60.6) | 2027 (54.3) | <.0001 |
| Moderate or heavy drinker (>155 drinks/y) | 1324 (29.2) | 1293 (34.6) | <.0001 |
Values are expressed as mean±SD for continuous factors and frequency (percentage) for categorical factors.
Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; BP, blood pressure; FBG, fasting blood glucose; LDL, low‐density lipoprotein; SPRINT, Systolic Blood Pressure Intervention Trial; TIA, transient ischemic attack; UACr, urine albumin to creatinine ratio.
Includes one participant with stage 5 chronic kidney disease (CKD; estimated glomerular filtration rate [eGFR] of 14.7 mL/min/1.73 m2).
Although SPRINT targeted a nondiabetic population, a small number of patients with diabetes mellitus (4.6%) were inadvertently included in the study, likely as a result of participants who were incidentally found to have increased fasting glucose levels (>126 mg/dL) without previously known history of diabetes mellitus. These patients with diabetes did not have a higher prevalence of AF at baseline. Furthermore, those with lower total cholesterol, LDL, and triglyceride values were more likely to have AF at baseline. While HDL was not associated with AF risk, the association between LDL and AF risk was age dependent such that in younger participants, a higher LDL level was associated with a lower prevalence of AF (age×LDL two‐way interaction, P=.044). Patients with AF at baseline were more likely to be treated with statins, renin‐angiotensin antagonists, and antithrombotic agents, likely reflecting the higher CVD burden in this subgroup. Patients with AF were also more likely to be moderate to heavy users of alcohol compared with those without AF at baseline. Use of niacin and fibrates was not different between AF and non‐AF groups.
In order to evaluate whether any particular component of MetS was associated with AF risk, we performed univariable logistic regression analyses (Table 3). Only the log‐transformed serum triglyceride levels showed a strong and persistent negative association with AF risk. The negative association between triglycerides and AF risk was not explained by differences in sex or age. While fasting serum glucose level was significantly associated with AF risk in the univariable analysis (Table 3), the association was no longer significant in the multivariable analysis (Table 4). The overall multivariable logistic regression analyses evaluating the association between AF risk and relevant clinical covariates are summarized in Table 4. Age, race, history of CVDs, decreased triglycerides, and albuminuria remained significantly associated with AF risk. Although AF was more prevalent in patients with higher PP at baseline (Table 1), the multivariable analysis unexpectedly demonstrated that PP was inversely associated with AF risk. The baseline systolic BP level remained unassociated with AF risk.
Table 3.
Univariable logistic regression model for the association between atrial fibrillation risk and metabolic syndrome and its components (N=8675)
| Covariates | ORa | 95% CI | P value |
|---|---|---|---|
| Metabolic syndromeb | 0.87 | 0.74–1.02 | .078 |
| Body mass index | 0.94 | 0.86–1.01 | .104 |
| Loge (triglycerides) | 0.87 | 0.80–0.94 | .0004 |
| Serum HDL cholesterol | 1.00 | 0.92–1.08 | .9 |
| Systolic BP | 0.96 | 0.89–1.04 | .4 |
| Serum fasting glucose | 1.08 | 1.00–1.16 | .037 |
| Diabetes mellitus | 0.82 | 0.59–1.17 | .2 |
The increments for odds ratios (ORs) for continuous factors: 5.8 kg/m2 for body mass index, 0.47 unit for Loge (triglycerides), 14.5 mg/dL for high‐density lipoprotein (HDL), 15.6 mm Hg for systolic blood pressure (BP), and 13.6 mg/dL for fasting glucose.
Metabolic syndrome was defined as having three or more of the following: (1) body mass index >30 kg/m2; (2) fasting triglycerides >150 mg/dL (or taking lipid‐lowering medication); (3) HDL cholesterol <50 mg/dL in women or <40 mg/dL in men; (4) BP ≥130/85 mm Hg or on treatment for hypertension; or (5) fasting glucose ≥100 mg/dL or diabetes mellitus.
Table 4.
Results of multivariable logistic regression model for atrial fibrillation risk and clinical covariates (N=8263)
| Factor | OR | 95% CI | P value |
|---|---|---|---|
| Age, y | 1.54 | 1.38–1.72 | <.0001 |
| Race | |||
| Black vs white | 0.46 | 0.36–0.59 | <.0001 |
| Hispanic vs white | 0.46 | 0.31–0.66 | |
| Other vs white | 1.25 | 0.68–2.13 | |
| Female sex | 0.88 | 0.71–1.09 | .3 |
| Body mass index | 1.08 | 0.98–1.20 | .13 |
| Current/former vs never smoker | 1.05 | 0.96–1.14 | .3 |
| Diabetes mellitus (history or FBG >126) | 1.01 | 0.65–1.55 | .9 |
| Metabolic syndrome | 0.97 | 0.76–1.24 | 1.0 |
| History of CAD | 2.25 | 1.86–2.73 | <.0001 |
| Congestive heart failure | 4.39 | 3.28–5.84 | <.0001 |
| Stroke/TIAa | 1.49 | 1.03–2.12 | .028 |
| Systolic BP | 1.05 | 0.93–1.19 | .4 |
| Pulse pressure | 0.83 | 0.73–0.96 | .009 |
| Loge (triglycerides) | 0.86 | 0.77–0.95 | .005 |
| LDL cholesterol | 0.98 | 0.88–1.08 | .7 |
| HDL cholesterol | 1.01 | 0.91–1.13 | .8 |
| eGFR | 1.00 | 0.91–1.10 | 1.0 |
| Loge (urine albumin‐creatinine ratio) | 1.17 | 1.08–1.26 | .0002 |
| Fasting serum glucose | 1.06 | 0.95–1.19 | .3 |
| ACEI/ARB use | 1.13 | 0.95–1.34 | .2 |
| Statin use | 1.04 | 0.86–1.28 | .7 |
| Niacin or fenofibrate use | 0.91 | 0.55–1.45 | .7 |
| Moderate or heavy drinker (>155 drinks/y) | 1.09 | 0.90–1.31 | .4 |
The increments for odds ratios (ORs) for continuous factors: 9.4 y for age, 20.6 mL/min/1.73 m2 for estimated glomerular filtration rate (eGFR), 1.18 unit for loge (urine albumin‐creatinine ratio), 5.8 kg/m2 for body mass index, 0.47 unit for loge (triglycerides), 14.5 mg/dL for high‐density lipoprotein (HDL), 15.6 mm Hg for systolic blood pressure (BP), 14.4 mm Hg for pulse pressure, 35 mg/dL for low‐density lipoprotein (LDL), and 13.6 mg/dL for fasting blood glucose (FBG).
Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; CAD, coronary artery disease. Bold values indicate significance.
Stroke, not transient ischemic attack (TIA), was exclusion for SPRINT (Systolic Blood Pressure Intervention Trial). The majority of the included cases refer to TIA, although a few patients had evidence of previous stroke on imaging studies after enrollment.
Given that SPRINT participants with AF had a higher prevalence of CVDs and statin therapy, additional multivariable analyses were stratified by baseline cardiovascular status and statin use to evaluate whether these differences could partly account for the lack of association between MetS and AF risk. In both cases, the lack of association between MetS and AF risk remained consistent. Because of the possibility that the differences in prevalence of MetS in the specific targeted high‐risk SPRINT subgroups may modulate the association between MetS and AF risk, we performed further analyses in which the odds ratio for AF risk in each targeted subgroup of SPRINT was assessed (Figure). In all targeted subgroups, the prevalence for MetS was increased, particularly in CKD (57%) and CVD (63%) subgroups. The prevalence of MetS was lowest in the senior subgroup (48%). Nonetheless, the lack of association between MetS and AF risk remained consistent across all subgroups.
Figure 1.
Metabolic syndrome as a risk factor for atrial fibrillation among the four high‐risk groups in SPRINT (Systolic Blood Pressure Intervention Trial). CKD indicates chronic kidney disease; CVD, cardiovascular disease
3.2. The association between AF risk and renal function
Decreased eGFR (<45 mL/min/1.73 m2) was associated with increased AF risk, as the prevalence of AF increased with worsening kidney function (Table 5). This positive association, however, was not statistically significant after adjustment for age and albuminuria. A separate logistic regression analysis limited to the SPRINT CKD subgroup also did not reveal any significant association of MetS with AF risk.
Table 5.
CKD groups and atrial fibrillation risk, unadjusted, adjusted for age, and adjusted for age and albuminuria
| CKD groups by eGFR, mL/min/1.73 m2 | Atrial fibrillation prevalence | OR (95% CI; unadjusted) | OR (95% CI; adjusted for age) | OR (95% CI; adjusted for age and albuminuria) | ||||
|---|---|---|---|---|---|---|---|---|
| CKD groups | No. | % | No. | % | ||||
| 1 | ≥90 | 1435 | 17.4 | 77 | 5.4 | 0.69 (0.53, 0.88) | 0.85 (0.66, 1.10) | 0.84 (0.65, 1.09) |
| 2 | 60–89 | 4473 | 54.1 | 340 | 7.6 | (reference) | (reference) | (reference) |
| 3A | 45–59 | 1569 | 19.0 | 165 | 10.5 | 1.43 (1.17, 1.73) | 1.17 (0.96, 1.43) | 1.14 (0.93, 1.33) |
| 3B | <45 | 786 | 9.5 | 95 | 12.1 | 1.67 (1.31, 2.12) | 1.32 (1.02, 1.63) | 1.14 (0.88, 1.47) |
Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; OR, odds ratio.
4. DISCUSSION
Dyslipidemia, obesity, and BMI have been strongly associated with AF risk in numerous studies, involving both the general population25, 26, 27 and patients with and without diabetes mellitus with underlying CVDs.28, 29, 30 The aim of our study was to investigate the cross‐sectional association between MetS and its components with prevalent AF in the entire SPRINT cohort and specifically in the CKD cohort. In contradiction to our hypothesis that MetS increases AF risk, MetS was not associated with AF in the SPRINT cohort. MetS was marginally associated with AF (P<.08) in univariable analysis, likely driven by fasting glucose. While the prevalence of AF increased with CKD, albuminuria was more strongly associated with AF risk in our study (Table 4). In multivariate analyses, significant predictors for AF were age, white race, preexisting CVDs, lower triglyceride levels, lower PP, and increased albuminuria.
The reasons for our unexpected results, including the inverse relationship of PP and triglycerides with AF risk, are unclear. Several factors may possibly contribute to our findings: (1) unmeasured confounders, (2) possible underestimation of the true AF rate in certain groups given heavy reliance on self‐report for AF diagnosis, or (3) a truly weak or no association of MetS (as commonly defined), with AF risk in the older SPRINT cohort.
SPRINT targeted a uniquely high‐risk population with older age and substantially higher prevalence of MetS at baseline (55% of the entire cohort) compared with the general population. The overall prevalence of MetS in the United States is 22% according to the third National Health and Nutrition Examination Survey (NHANES III).31 The prevalence for MetS was 43.5% for patients 60 to 69 years and 42% for patients older than 70 years in NHANES, suggesting that our SPRINT cohort had a significantly higher prevalence of MetS ranging between 48% (the senior subgroup) and 63% (the CVD subgroup). Given the high prevalence of metabolic derangement in our study population, it may have been difficult to isolate the effect of MetS on AF risk. Moreover, women were underrepresented in our study, accounting for 36% of the total cohort and 28% of AF cases. It is unknown whether the components of MetS confer differential effects on AF risk in women compared with men.
Age may also have important implications for the distribution of lipid measurements. Total cholesterol level increases with age in relatively younger populations but decreases in older individuals beyond age 60 years.32 While men tend to have higher triglyceride levels than women through the younger adult years, triglyceride levels increase in postmenopausal women, with higher levels than men from 60 years onward.33 Taken together, the combined effects of increased age and greater male participation in SPRINT could partly explain the negative relationship between triglycerides and AF risk. In our analyses, however, the negative association between triglyceride levels and AF risk was not explained by age or male sex. An important confounder may be the greater use of statins in patients with AF, resulting in lower triglyceride values and thus contributing to lack of association with MetS. The additional multivariable analysis stratified by baseline statin use, however, still demonstrated consistent lack of association.
Age may also have a unique effect on the relationship between PP and AF risk. PP serves as a surrogate marker of arterial stiffness and vascular dysfunction and has been known to contribute to AF development.34,35 Our finding that PP is inversely associated with AF risk is thus unexpected and is difficult to explain. Interestingly, a large cross‐sectional study of data from 25,109 participants of the REGARDS (Reasons for Geographic and Racial Differences in Stroke)36 study have reported that the association between PP and AF risk differed by age, with those older (≥75 years) actually having a lower risk of AF with higher PP, similar to our results. It is unclear why age may have a modulating effect on the relationship between PP and AF. It remains unknown whether mechanisms other than vascular stiffness underlie the association between PP and AF risk and if those factors are differentially affected by age.
In addition, it is unclear whether the high representation of blacks (30.4%) in SPRINT could have affected the study results. Blacks are less likely than whites to have either elevated triglyceride or low HDL levels.37 Blacks only accounted for 4.4% of all AF cases, however, and it is doubtful that black race had a significant influence on our negative results. While the inverse association between dyslipidemia and AF risk may reflect unmeasured confounders in our study, various investigators have questioned whether there may be a physiologic basis, including thyroid dysfunction and inflammation.4,32 While thyroid hormones stimulate the synthesis of cholesterol by inducing 5‐hydroxy‐3‐methylglutaryl‐coenzyme A reductase, they also upregulate the hepatic catabolism of cholesterol, leading to a reduction of LDL levels. Given the inverse relationship between thyroid hormone and total cholesterol and LDL levels39 and the established role of hyperthyroidism (both clinical and subclinical) in the development of AF,40 it is plausible that low cholesterol levels are associated with higher AF risk, particularly in the elderly. Unfortunately, SPRINT did not assess thyroid function as a part of the original study, and none of the large trials assessing the role of dyslipidemia and AF risk have investigated the role of thyroid function.
Moreover, inflammation may partially explain the inverse relationship between dyslipidemia and AF risk. One study has suggested that LDL was inversely associated with C‐reactive protein level, a known risk factor for AF.41 Given that inflammation is associated with pathogenesis and development of AF,42,43 low LDL or cholesterol levels may be linked to increased AF risk through increased inflammation. Our study did not measure markers of inflammation, and thus we are unable to investigate this possibility. Although we lack direct measures of inflammation, it is curious that albuminuria was a stronger predictor of AF risk than eGFR in our study. This may in part be attributable to the fact that the majority of the SPRINT population had mild CKD, with nearly 54% of the cohort having eGFR >60 mL/min/1.73 m2 and only 9.64% having eGFR <45 mL/min/1.73 m2. A more intriguing possibility is that albuminuria not only reflects renal function but a more generalized vascular endothelial dysfunction44 and inflammation, as demonstrated by a significant correlation with C‐reactive protein levels.45 It is thus plausible that the association between albuminuria and AF risk may represent the importance of inflammation as the pathogenic factor.
Alternatively, MetS and its components may not play an important role in AF risk in an older, hypertensive population at high risk for CVD. A Taiwanese study involving 3775 patients with important shared features with those in our study (similar mean age 67.4±6.9 years and high prevalence of MetS and hypertension) also suggested a lack of association between MetS and AF risk in elderly patients.38
5. STUDY STRENGTHS
The strengths of our study include a large sample size and the diversity of the population (including a large proportion of patients ≥75 years and a significant proportion of blacks). Conversely, our study is limited by the stringent inclusion criteria to target high‐risk groups, diminishing generalizability to populations not included in the study, such as those with diabetes mellitus, those with prior stroke, and those younger than 50 years. Our definition for MetS was inherently limited by the available SPRINT data, which may have contributed to our negative results. Most of our AF cases were derived from self‐report, which may underestimate the true prevalence of disease. It is also unclear whether the underrepresentation of women may have impacted our study results.
6. CONCLUSIONS
AF remains a complex clinical entity with diverse causes. MetS is not independently associated with AF in older populations with hypertension.
SPRINT ACKNOWLEDGMENT
We also acknowledge support from the following Clinical and Translational Science Awards funded by National Center for Advancing Translational Sciences: Case Western Reserve University: UL1TR000439; Ohio State University: UL1RR025755; University of Pennsylvania: UL1RR024134 and UL1TR000003; Boston: UL1RR025771; Stanford: UL1TR000093; Tufts: UL1RR025752, UL1TR000073, and UL1TR001064; University of Illinois: UL1TR000050; University of Pittsburgh: UL1TR000005; UT Southwestern: 9U54TR000017‐06; University of Utah: UL1TR000105‐05; Vanderbilt University: UL1 TR000445; George Washington University: UL1TR000075; University of CA, Davis: UL1 TR000002; University of Florida: UL1 TR000064; University of Michigan: UL1TR000433; and Tulane University: P30GM103337 COBRE Award NIGMS.
CONFLICT OF INTEREST
None.
ACKNOWLEDGMENTS
SPRINT is funded with federal funds from the National Institutes of Health (NIH), including the National Heart, Lung, and Blood Institute (NHLBI); the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK); the National Institute on Aging (NIA); and the National Institute of Neurological Disorders and Stroke (NINDS), under contract numbers HHSN268200900040C, HHSN268200900046C, HHSN268200900047C, HHSN268200900048C, HHSN268200900049C, and Inter‐Agency Agreement Number A‐HL‐13‐002‐001. It is also supported in part with resources and use of facilities through the Department of Veterans Affairs. SPRINT investigators acknowledge the contribution of study medications (azilsartan and azilsartan combined with chlorthalidone) from Takeda Pharmaceuticals International, Inc. All components of SPRINT protocol were designed and implemented by the investigators. The investigative team collected, analyzed, and interpreted the data. All aspects of manuscript writing and revision were performed by the coauthors. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH, the US Department of Veterans Affairs, or the US Government. For a full list of contributors to SPRINT, please see the supplementary acknowledgment list:
ME Cho, TE Craven, AK Cheung, et al. The association between insulin resistance and atrial fibrillation: A cross‐sectional analysis from SPRINT (Systolic Blood Pressure Intervention Trial). J Clin Hypertens. 2017;19:1152–1161. 10.1111/jch.13062
REFERENCES
- 1. Bansal N, Fan D, Hsu CY, et al. Incident atrial fibrillation and risk of death in adults with chronic kidney disease. J Am Heart Assoc. 2014;3:e001303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Bansal N, Fan D, Hsu CY, et al. Incident atrial fibrillation and risk of end‐stage renal disease in adults with chronic kidney disease. Circulation. 2013;127:569‐574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ananthapanyasut W, Napan S, Rudolph EH, et al. Prevalence of atrial fibrillation and its predictors in nondialysis patients with chronic kidney disease. Clin J Am Soc Nephrol. 2010;5:173‐181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tadic M, Ivanovic B, Cuspidi C. What do we currently know about metabolic syndrome and atrial fibrillation? Clin Cardiol. 2013;36:654‐662. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Watanabe H, Tanabe N, Watanabe T, et al. Metabolic syndrome and risk of development of atrial fibrillation: the Niigata preventive medicine study. Circulation. 2008;117:1255‐1260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Huxley RR, Alonso A, Lopez FL, et al. Type 2 diabetes, glucose homeostasis and incident atrial fibrillation: the Atherosclerosis Risk in Communities study. Heart. 2012;98:133‐138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Dublin S, Glazer NL, Smith NL, et al. Diabetes mellitus, glycemic control, and risk of atrial fibrillation. J Gen Intern Med. 2010;25:853‐858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Johansen OE, Brustad E, Enger S, et al. Prevalence of abnormal glucose metabolism in atrial fibrillation: a case control study in 75‐year old subjects. Cardiovasc Diabetol. 2008;7:28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Latini R, Staszewsky L, Sun JL, et al. Incidence of atrial fibrillation in a population with impaired glucose tolerance: the contribution of glucose metabolism and other risk factors. A post hoc analysis of the Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research trial. Am Heart J. 2013;166:935‐940.e931. [DOI] [PubMed] [Google Scholar]
- 10. Chao TF, Suenari K, Chang SL, et al. Atrial substrate properties and outcome of catheter ablation in patients with paroxysmal atrial fibrillation associated with diabetes mellitus or impaired fasting glucose. Am J Cardiol. 2010;106:1615‐1620. [DOI] [PubMed] [Google Scholar]
- 11. Gunarathne A, Patel JV, Kausar S, et al. Glycemic status underlies increased arterial stiffness and impaired endothelial function in migrant South Asian stroke survivors compared to European Caucasians: pathophysiological insights from the West Birmingham Stroke Project. Stroke. 2009;40:2298‐2306. [DOI] [PubMed] [Google Scholar]
- 12. Lastra G, Dhuper S, Johnson MS, et al. Salt, aldosterone, and insulin resistance: impact on the cardiovascular system. Nat Rev Cardiol. 2010;7:577‐584. [DOI] [PubMed] [Google Scholar]
- 13. Soliman EZ, Prineas RJ, Go AS, et al. Chronic kidney disease and prevalent atrial fibrillation: the Chronic Renal Insufficiency Cohort (CRIC). Am Heart J. 2010;159:1102‐1107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Beltran‐Sanchez H, Harhay MO, Harhay MM, et al. Prevalence and trends of metabolic syndrome in the adult U.S. population, 1999–2010. J Am Coll Cardiol. 2013;62:697‐703. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Basturk T, Unsal A. What is the frequency of carbohydrate metabolism disorder in CKD? J Ren Care. 2012;38:15‐21. [DOI] [PubMed] [Google Scholar]
- 16. Townsend RR, Anderson AH, Chen J, et al. Metabolic syndrome, components, and cardiovascular disease prevalence in chronic kidney disease: findings from the Chronic Renal Insufficiency Cohort (CRIC) Study. Am J Nephrol. 2011;33:477‐484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wang CJ, Bao XR, Du GW, et al. Effects of insulin resistance on left ventricular hypertrophy in patients with CKD stage 1‐3. Int Urol Nephrol. 2014;46:1609‐1617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Shinohara K, Shoji T, Emoto M, et al. Insulin resistance as an independent predictor of cardiovascular mortality in patients with end‐stage renal disease. J Am Soc Nephrol. 2002;13:1894‐1900. [DOI] [PubMed] [Google Scholar]
- 19. Marott SC, Nielsen SF, Benn M, et al. Antihypertensive treatment and risk of atrial fibrillation: a nationwide study. Eur Heart J. 2014;35:1205‐1214. [DOI] [PubMed] [Google Scholar]
- 20. Vermes E, Tardif JC, Bourassa MG, et al. Enalapril decreases the incidence of atrial fibrillation in patients with left ventricular dysfunction: insight from the studies of left ventricular dysfunction (SOLVD) trials. Circulation. 2003;107:2926‐2931. [DOI] [PubMed] [Google Scholar]
- 21. Group SR, Wright JT Jr, Williamson JD, et al. A randomized trial of intensive versus standard blood‐pressure control. N Engl J Med. 2015;373:2103‐2116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Ambrosius WT, Sink KM, Foy CG, et al. The design and rationale of a multicenter clinical trial comparing two strategies for control of systolic blood pressure: the systolic blood pressure intervention trial (SPRINT). Clin Trials. 2014;11:532‐546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Matthews DR, Hosker JP, Rudenski AS, et al. Homeostasis model assessment: insulin resistance and beta‐cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412‐419. [DOI] [PubMed] [Google Scholar]
- 24. Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults . Executive summary of the third report of the national cholesterol education program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel III). JAMA. 2001;285:2486‐2497. [DOI] [PubMed] [Google Scholar]
- 25. Chamberlain AM, Agarwal SK, Ambrose M, et al. Metabolic syndrome and incidence of atrial fibrillation among blacks and whites in the Atherosclerosis Risk in Communities (ARIC) Study. Am Heart J. 2010;159:850‐856. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Tentolouris N, Liatis S, Katsilambros N. Sympathetic system activity in obesity and metabolic syndrome. Ann N Y Acad Sci. 2006;1083:129‐152. [DOI] [PubMed] [Google Scholar]
- 27. Wilhelmsen L, Rosengren A, Lappas G. Hospitalizations for atrial fibrillation in the general male population: morbidity and risk factors. J Intern Med. 2001;250:382‐389. [DOI] [PubMed] [Google Scholar]
- 28. Frost L, Hune LJ, Vestergaard P. Overweight and obesity as risk factors for atrial fibrillation or flutter: the Danish diet, cancer, and health study. Am J Med. 2005;118:489‐495. [DOI] [PubMed] [Google Scholar]
- 29. Verdecchia P, Dagenais G, Healey J, et al. Blood pressure and other determinants of new‐onset atrial fibrillation in patients at high cardiovascular risk in the ongoing telmisartan alone and in combination with ramipril global endpoint trial/telmisartan randomized assessment study in ACE intolerant subjects with cardiovascular disease studies. J Hypertens. 2012;30:1004‐1014. [DOI] [PubMed] [Google Scholar]
- 30. Wanahita N, Messerli FH, Bangalore S, et al. Atrial fibrillation and obesity–results of a meta‐analysis. Am Heart J. 2008;155:310‐315. [DOI] [PubMed] [Google Scholar]
- 31. Ford ES, Giles WH, Dietz WH. Prevalence of the metabolic syndrome among US adults: findings from the third national health and nutrition examination survey. JAMA. 2002;287:356‐359. [DOI] [PubMed] [Google Scholar]
- 32. Suzuki S. “Cholesterol paradox” in atrial fibrillation. Circ J. 2011;75:2749‐2750. [DOI] [PubMed] [Google Scholar]
- 33. Carroll MD, Lacher DA, Sorlie PD, et al. Trends in serum lipids and lipoproteins of adults, 1960‐2002. JAMA. 2005;294:1773‐1781. [DOI] [PubMed] [Google Scholar]
- 34. Darne B, Girerd X, Safar M, et al. Pulsatile versus steady component of blood pressure: a cross‐sectional analysis and a prospective analysis on cardiovascular mortality. Hypertension. 1989;13:392‐400. [DOI] [PubMed] [Google Scholar]
- 35. Mitchell GF, Vasan RS, Keyes MJ, et al. Pulse pressure and risk of new‐onset atrial fibrillation. JAMA. 2007;297:709‐715. [DOI] [PubMed] [Google Scholar]
- 36. Ghazi L, Safford MM, Khodneva Y, et al. Gender, race, age, and regional differences in the association of pulse pressure with atrial fibrillation: the reasons for geographic and racial differences in stroke study. J Am Soc Hypertens. 2016;10:625‐632.e621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Sumner AE. Ethnic differences in triglyceride levels and high‐density lipoprotein lead to underdiagnosis of the metabolic syndrome in black children and adults. J Pediatr. 2009;155:S7.e7–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Hu YF, Chen SA, Yeh HI. Metabolic syndrome does not impose greater atrial fibrillation risk in elderly hypertensive patients. Acta Cardiol. 2010;65:653‐659. [DOI] [PubMed] [Google Scholar]
- 39. Asvold BO, Vatten LJ, Nilsen TI, et al. The association between TSH within the reference range and serum lipid concentrations in a population‐based study. The HUNT Study. Eur J Endocrinol. 2007;156:181‐186. [DOI] [PubMed] [Google Scholar]
- 40. Gencer B, Collet TH, Virgini V, et al. Subclinical thyroid dysfunction and cardiovascular outcomes among prospective cohort studies. Endocr Metab Immune Disord Drug Targets. 2013;13:4‐12. [DOI] [PubMed] [Google Scholar]
- 41. Folsom AR, Pankow JS, Tracy RP, et al. Association of C‐reactive protein with markers of prevalent atherosclerotic disease. Am J Cardiol. 2001;88:112‐117. [DOI] [PubMed] [Google Scholar]
- 42. Aviles RJ, Martin DO, Apperson‐Hansen C, et al. Inflammation as a risk factor for atrial fibrillation. Circulation. 2003;108:3006‐3010. [DOI] [PubMed] [Google Scholar]
- 43. Chung MK, Martin DO, Sprecher D, et al. C‐reactive protein elevation in patients with atrial arrhythmias: inflammatory mechanisms and persistence of atrial fibrillation. Circulation. 2001;104:2886‐2891. [DOI] [PubMed] [Google Scholar]
- 44. Deckert T, Feldt‐Rasmussen B, Borch‐Johnsen K, et al. Albuminuria reflects widespread vascular damage. The Steno hypothesis. Diabetologia. 1989;32:219‐226. [DOI] [PubMed] [Google Scholar]
- 45. Agrawal V, Marinescu V, Agarwal M, et al. Cardiovascular implications of proteinuria: an indicator of chronic kidney disease. Nat Rev Cardiol. 2009;6:301‐311. [DOI] [PubMed] [Google Scholar]