The association of the metabolic syndrome with QTc interval in NHANES III

  • Mohammed F. Faramawi
  • Rachel P. Wildman
  • Jeanette Gustat
  • Janet Rice
  • Mohammed Y. Abdul Kareem
CARDIOVASCULAR DISEASE

DOI: 10.1007/s10654-008-9252-4

Cite this article as:
Faramawi, M.F., Wildman, R.P., Gustat, J. et al. Eur J Epidemiol (2008) 23: 459. doi:10.1007/s10654-008-9252-4

Abstract

Purpose To examine the relationship of metabolic syndrome with corrected QT interval duration. Methods In a cross-sectional analysis of NHANES III, a representative sample of the adult US population, 3,495 individuals aged ≥40 years were categorized as having metabolic syndrome and not having metabolic syndrome as defined by ATP III. QT interval was measured from the standard 12-lead electrocardiogram. Results A weighted multi-linear regression analysis adjusted for age, gender, serum calcium, and potassium showed that metabolic syndrome was significantly associated with a corrected QT interval in milliseconds. The adjusted Beta coefficient and its corresponding standard error were (4.48 milliseconds, 0.95 milliseconds, P < 0.01). Conclusion These data indicate that metabolic syndrome is independently associated with a corrected QT interval duration. This study calls for careful ECG monitoring among persons with metabolic syndrome for early detection of a long corrected QT interval in order to prevent severe and often fatal arrhythmias.

Keywords

Metabolic syndrome Cross-sectional studies Electrocardiography 

Introduction

The overall prevalence of the metabolic syndrome (MetS) in the United States is >20% in men and women over the age of 20 years and >40% in men and women over the age of 60 years [1]. A number of studies have shown that MetS is associated with cardiovascular diseases such as coronary heart disease (CHD) and peripheral vascular disease [2, 3, 4]. However, few studies have examined the association of this syndrome with electrocardiographic subclinical risk factors such as a long corrected QT interval (QTc) [5]. Epidemiological studies have shown that prolongation of the QTc interval has been associated with an increased risk of sudden death [6, 7] coronary heart disease, and all cause mortality in a broad range of clinical populations as well as in healthy subjects in population based studies [8, 9, 10, 11]. As little is known about the association of MetS with QTc interval duration in population-based studies, this analysis aimed to examine the relationship of MetS with QTc interval in a nationally representative sample of the adult US population.

Methods

Study population

Between 1988 and 1994, the National Center for Health Statistics conducted the third National Healthy and Nutrition Examination Survey (NHANES III). This cross-sectional study consisted of a multistage stratified clustered probability sample of the US civilian noninstitutionalized population. Because NHANES III is based on a complex multistage probability sample design, appropriate probability sampling weights were assigned to produce unbiased population estimates. The sampling weights incorporate the differential probabilities of selection and adjust for non-coverage and non-response. Included in the present analysis were NHANES III participants aged 40 years and older, who participated in an interview conducted at home and who had data available from their physical, electorcardiogrphic and laboratory examinations.

NHANES III consisted of a standardized questionnaire administered in the home by a trained interviewer followed by a detailed physical examination at a Mobile Examination Center (MEC). Self-reported data collected at the home interview relevant to the current analysis included demographics (age, race and gender), a history of diabetes, history of high blood pressure (hypertension), heart disease and thyroid illness. Hypertension status was determined by self-reporting of hypertension and measuring the resting blood pressure levels. Participants were coded as having hypertension if they answered yes to one of the following questions from the Household Adult Questionnaire: “Have you ever been told by a doctor or other health professional that you had hypertension, also called high blood pressure?” “Because of your (high blood pressure/hypertension), have you ever been told by a doctor or other health professional to take prescribed medicine?.” Up to six blood pressure measurements were taken on two occasions according to a standard protocol; the first set of three blood pressures was measured in the home by a trained and certified interviewer and a physician obtained the second set during the medical examination. The average of all available blood pressure measurements was used for data analysis.

Presence of heart disease was based on the interview questions: “Has a doctor ever told you that you had a heart attack?” “Has a doctor ever told you that you had congestive heart failure?” “Have you been hospitalized for a heart problem (i.e., heart attack, angina or chest pain, congestive heart failure.” Thyroid illness was determined on the basis of affirmative answers to the following questions “Has the doctor ever told you had goiter?” “Has the doctor ever told you had thyroid disease?” “Do you still have thyroid disease?”

Information about usage of QT interval prolonging medications such as methadone, anti-infective agents (Macrolides, sparfloxacin, chloroquine) antipsychotic drugs or anti-manic drugs such as (Lithium, chlorpromazine, haloperidol, mesoridazole, thioridazone and pimozide) diuretics (indapamide), Anti-arrhythmic (disopyramide, quinidine, procainamide,sotalol and amiodarone) and gastrointestinal prokinetic drugs such as cisapride [12, 13, 14] in the past 30 days was collected during the household interview. The previously listed drugs were recorded and their containers were examined by the interviewer. During the physical examination at the Mobile Examination Center a 24-h dietary recall was administered, which assessed the amount of alcohol consumed during the previous day. The data collected were used to calculate the daily intake of alcohol in grams. The daily alcohol consumption was categorized into ≥1 g/day or <1 g/day.

Measurements of the MetS components

During the visit to the mobile examination center, fasting venous blood specimen was drawn from each participant according to a standardized protocol [15]. Plasma glucose was measured at the University of Missouri Diabetes Diagnostic Laboratory using a hexokinase enzymatic method [16]. Fasting serum insulin, total cholesterol, High density lipoprotein cholesterol (HDL), low density lipoprotein cholesterol (LDL), and triglycerides were measured at other centralized laboratories [15, 17]. At the mobile examination center, serum and plasma were separated from the specimens within 1 h of collection, transported on dry ice to the participating laboratories, and stored refrigerated or frozen, as appropriate, until analysis. Serum sodium and potassium assays were carried out by the participating laboratories.

The revised ATP III guidelines were used to define the MetS [18] as the presence of three or more of the following: (1) A systolic blood pressure ≥130 and/or diastolic blood pressure ≥85 mmHg, (2) HDL <40 mg/dl for men and <50 mg/dl for women, (3) Triglycerides ≥150 mg/dl, (4) Fasting blood glucose ≥100 mg/dl, (5) Waist circumference >102 cm for men and >88 cm for women. Cotinine concentration in the blood, a metabolite of nicotine, was used as a biomarker to classify participants into current and non-current smokers. Current smokers were defined as those who had cotinine levels >15 ng/ml, while those with serum cotinine ≤15 ng/ml were classified as not current smokers [19, 20].

Measurement of the QT interval and other electrocardiographic components

A standard 12-lead resting ECG was performed on all men and women aged 40 years or over using a Marquette MAC 12 (Marquette Medical System, Inc., Milwaukee, WI, USA). ECG data were recorded with eight independent components of the 12 standard leads simultaneously and were sampled at 250 samples per second per channel. Using the Dalhousie ECG Analysis Program, a representative P-QRS-T cycle was derived by selective averaging. The corrected QT interval (QTc), which adjusts for the effects of heart rate on QT duration, was the outcome variable. Multiple QT correction formulas exist to account for heart rate. In order to determine the most appropriate correction formula, QTc by heart rate was assessed with univariable linear regression models [21]. As QTc corrects for heart rate, no relationship should be observed between heart rate and QTc after correction has been performed. An association was noticed with the Bazett’s equation but not with the Fridericia’s equation. Therefore Fridericia’s formula was assigned for QT interval correction [22].

Statistical analysis

Of 8,561 subjects who underwent ECG testing, 194 were removed as a result of missing data about QT interval and heart rate. Other participants were excluded because they had missing information about the key variables of metabolic syndrome (n = 645), daily alcohol consumption (n = 258) or serum calcium (n = 119). Also those who were on QT interval prolonging drugs (n = 156) or had thyroid illness (n = 390), and cardiovascular disease or ECG with atrioventricular conduction abnormalities (n = 944) were excluded. Heart disease was defined as self-reported history of heart attack, heart failure and myocardial infarction. Finally, 2,460 participants were excluded because they did not fast for 8 h or more. This left 3,495 participants for the current analyses.

All analyses were performed using Stata version 9 which took into account the complex sampling techniques used in NHANES III. Observations were weighted using weights calculated for that purpose by the National Center for Health Statistics to reflect the general United States population. These weights were also designed to adjust for biases attributable to non-response.

Basic descriptive statistics, including means and percentages were used to characterize the study participants. Comparison of the continuous variables means by MetS status was made by conducting unpaired t-tests, while comparison of the categorical variables across the two groups was made by the χ2 test.

The QTc duration was normally distributed. Bivariate linear regression models were conducted to examine the relationship of the different potential confounders with QTc interval separately. To be considered as a confounder, the variable had to be associated with both MetS and QTc interval. Variables considered as possible confounders included age, gender, race, smoking, alcohol consumption, serum calcium and potassium.

Additionally, multivariate models testing for interaction terms (effect modification) between MetS with gender, age and race separately were conducted. To identify the independent contribution of each metabolic syndrome component with QTc interval duration, a multivariable linear regression analysis was conducted. In this analysis the different metabolic syndrome components (elevated systolic blood pressure, elevated diastolic blood pressure, elevated blood triglycerides, increased waist circumference, impaired fasting blood glucose and low HDL) were included.

Results

Participants with MetS were more often males and less often alcohol consumers (Table 1). Individuals with MetS were older than those who did not have the syndrome (Table 1). Smoking and race were not associated with MetS. In regards to the MetS components, the mean values of systolic blood pressure, diastolic blood pressure, triglycerides concentration, fasting blood glucose level, and waist circumference were higher in MetS group while HDL concentration was lower in the same group (Table 1). Serum calcium and potassium levels were higher in the MetS group (Table 1).
Table 1

Demographic characteristics and habits, serum electrolytes, QTc duration in milliseconds and MetS components by MetS status

Variablesa

MetS

P value

Absent

Present

N = 1,944

N = 1,551

Age (years)

53.02 (0.50)

57.94 (0.55)

<0.01

Male n (%)

979 (45.39)

750 (52.20)

<0.01

Race

  White n (%)

1,392 (86.65)

1,170 (88.91)

0.06

  Black n (%)

487 (08.93)

340 (08.59)

 

  Others n (%)

65 (04.42)

41 (02.50)

 

Current smoker n (%)

589 (29.07)

393 (27.43)

0.32

Alcohol consumption ≥1 g/day n (%)

437 (25.64)

249.00 (16.64)

<0.01

QTc interval (milliseconds)

416.93 (0.81)

423.50 (0.70)

<0.01

Systolic blood pressure (mmHg)

120.12 (0.50)

136.58 (0.62)

<0.01

Diastolic blood pressure (mmHg)

73.20 (0.21)

80.47 (0.36)

<0.01

HDL (mg/dl)

55.06 (0.50)

43.80 (0.70)

<0.01

Triglycerides (mg/dl)

89.32 (3.62)

152.72 (7.50)

<0.01

Fasting blood glucose (mg/dl)

85.45 (3.45)

103.67 (2.33)

<0.01

Waist circumference (cm)

89.83 (0.32)

103.28 (0.43)

<0.01

Serum potassium (mg/dl)

4. 01 (0.01)

4.70 (0.02)

<0.01

Serum calcium (mg/dl)

9.78 (0.03)

10.65(0.02)

<0.01

aVariable values Mean (SE) or Number (%)

The overall mean QTc was 419.73 milliseconds and the standard error was 0.56. Individuals with MetS had a longer QTc interval than those without MetS (423.50 milliseconds vs. 416.93 milliseconds, respectively, P < 0.01) (Table 1). In regards to QTc, it was longer in females and individuals with MetS (Table 2). QTc was positively associated with age while negatively associated with serum calcium and potassium (Table 2). Smoking and race were not statistically associated with MetS or QTc interval.
Table 2

The relationship of MetS, personal characteristics and habits, serum electrolytes with QTc interval duration in milliseconds

 

Unadjusted analysis

N = 3,495

B

SE

P

Metabolic syndrome

  No (n = 1,944)

Reference

  

  Yes (n = 1,551)

6.57

(1.10)

<0.01

  Age (years)

0.40

(0.03)

<0.01

Gender

  Male (n = 1,766)

Reference

  

  Female (n = 1,729)

9.04

(0.89)

<0.01

Race

  White (n = 2,562)

Reference

  

  Black (n = 827)

1.31

(0.98)

0.20

  Other (n = 106)

2.90

(2.38)

0.09

Smoking

  Current(n = 2,513)

Reference

  

  Non-current (n = 982)

−1.60

(1.07)

0.20

Alcohol consumption (g/day)

  <1 g/day (n = 2,809)

Reference

  

  ≥1 g/day (n = 686)

0.35

(1.16)

0.76

Serum Potassium (mg/dl)

−6.92

(1.54)

<0.01

Serum Calcium (mg/dl)

−3.62

(1.13)

<0.01

Variables which were associated with both MetS and QTc interval included gender, serum calcium and potassium. Smoking and race were not statistically associated with MetS or QTc interval, while alcohol consumption status had a significant association with MetS but not with QTc interval. Therefore, smoking, alcohol consumption and race were not considered as confounders. The final model included in addition to metabolic syndrome as an independent variable, age, gender, serum calcium and potassium as confounders. An alternative model was run adjusting for all potential confounders including smoking, alcohol consumption and race, and results were similar. Therefore, results from the more parsimonious model are presented here. The obtained Beta coefficients were considered statistically significant if their P values were <0.05.

After adjusting for age, gender, serum potassium, and serum calcium, QTc interval duration remained significantly longer in the MetS group (Table 3). When the components of the metabolic syndrome were considered, elevated systolic blood pressure, increased waist circumference, impaired fasting blood glucose and low HDL were significant predictors for QTc interval prolongation (Table 4). Interaction terms between MetS and age, race, and gender were not significant (P > 0.05 for all).
Table 3

The relationship of MetS with QTc interval duration in milliseconds after adjusting for the confounders

 

Adjusted analysis

N = 3,495

B

SE

P

Metabolic syndrome

  Absent (n = 1,944)

Reference

  

  Present (n = 1,551)

4.48

(0.95)

<0.01

  Age (years)

0.40

(0.03)

<0.01

Gender

  Male (n = 1,766)

Reference

  

  Female (n = 1,729)

7.97

(0.97)

<0.01

Serum Potassium (mg/dl)

6.28

(1.48)

<0.01

Serum Calcium (mg/dl)

3.14

(1.10)

<0.01

Adjusted for age, gender, serum potassium and calcium

Table 4

The relationship of MetS components with QTc duration in milliseconds

 

Unadjusted analysis

Adjusted analysis

N = 3,495

N = 3,495

B

SE

P

B

SE

P

Increased waist circumference

  Normal (n = 1,808)

Reference

  

Reference

  

  Increased (n = 1,687)

5.30

(0.94)

<0.01

4.70

(0.83)

<0.01

Low HDL

  Normal (n = 2,196)

Reference

  

Reference

  

  Low (n = 1,299)

2.84

(0.04)

<0.01

2.22

(0.85)

0.02

Systolic blood pressure

  Normal (n = 1,872)

Reference

  

Reference

  

High (n = 1,623)

8.14

(1.07)

<0.01

8.23

(1.31)

<0.01

High diastolic blood pressure

  Normal (n = 2,810)

Reference

  

Reference

  

  High (n = 685)

3.22

(1.50)

0.04

0.70

(1.67)

0.70

Impaired glucose tolerance

  Normal (n = 1,973)

Reference

  

Reference

  

  Impaired (n = 1,522)

4.02

(1.20)

<0.01

3.20

(1.14)

<0.01

High triglycerides

  Normal (n = 2,254)

Reference

  

Reference

  

  High (n = 1,241)

0.43

(1.19)

0.78

0.32

(1.21)

0.80

Elevated systolic blood pressure if values are ≥130. Elevated diastolic blood pressure if the values are ≥85 mmHg. Low HDL if the values are <40 mg/dl for men and <50 mg/dl for women. High triglycerides if the values ≥150 mg/dl. Impaired fasting blood glucose if the values ≥100 mg/dl. Increased weight circumference if the waist circumference >102 cm for men and >88 cm for women

Discussion

This study found a significant relationship between MetS and QTc interval. Participants with MetS had a longer QT interval duration than those who did not have the syndrome. Although participants with MetS had significantly higher values of QTc than those who did not have the syndrome (P < 0.001), the absolute mean values were not abnormal in this study. However, QT interval prolongation, even within the normal range, has been reported to be associated with adverse cardiovascular outcomes in the epidemiological studies [5, 23, 24, 25].

The present analysis showed that elevated systolic blood pressure, increased waist circumference, impaired fasting blood glucose and low HDL were statistically associated with QTc prolongation. High systolic blood pressure, dyslipidemia, increased waist circumference and impaired fasting blood glucose are associated with impaired nitric oxide availability which leads to pathological changes in the blood vessels (angiopathy) [26] and subsequent development of myocardial disease.

Hyperglycemia induces glucose toxicity which stimulates oxidative stress leading to endothelial dysfunction, Hypertension may lead to the elevation of soluble adhesion molecules [27] and decreases in the nitric oxide concentration resulting in endothelial dysfunction and damage [28]. Dyslipidemia may induce endothelial dysfunction by increasing oxidative stress. While the increase of waist circumference may lead to endothelial dysfunction via modulation of the secretion of adipocyte mediators, such as fatty acids, tumor necrosis factor-alpha and adiponectin. Endothelial damage and dysfunction will eventually lead to abnormal angiogenesis and decrease of blood flow to the different body organs such as the heart and kidney. Therefore, we hypothesize that participants with MetS suffer from pathological changes in their coronary arteries which lead to impairment of the blood flow to the myocardial muscle and development of subclinical or clinical myocardial disease. Myocardial disease may lead to prolongation of ventricular repolarization which appears as a longer QTc [29, 30] in individuals with MetS. Longer QTc interval predisposes patients with MetS to developing adverse cardiovascular outcomes such as torsade de pointes (ventricular arrhythmias).

The results of this study among a larger, representative sample of the US adult population, including data from multiple race-ethnic groups, confirm those reported by Soydinc et al. among a small Turkish population [5]. These findings have important clinical and public health implications due to the high prevalence of MetS in the United States [1]. The overall prevalence of the MetS is >20% in men and women over the age of 20 years and >40% in men and women over the age of 60 years [1]. Between 1988–1994 and 1999–2000, a significant increase in the prevalence of the MetS occurred among US adults aged ≥20 years [31]. Epidemiological studies have shown that prolongation of the QTc interval has been associated with an increased risk of sudden death due to fatal ventricular arrhythmia [6], coronary heart disease and all cause mortality in a broad range of clinical populations as well as in healthy subjects [8, 9, 10, 11]. In these studies, the relative risks for all cause morality ranged from 1.28 to 2.30, while the relative risks of cardiac mortality ranged from 1.36 to 2.51 [8, 9, 10, 11]. Therefore, careful monitoring of MetS patients to detect ventricular arrhythmias is warranted [32, 33].

This study has several strengths. Careful measurement of MetS (independent variable) and QTc interval duration (dependant variable) allowed precise estimation of the association. Also, this is the first population-based study to report a relationship between the MetS, defined by the ATP III guidelines, and the risk of having a longer QT interval. Nevertheless, the results of this study should be interpreted with caution. Data about heart disease and thyroid diseases were acquired from patients’ responses and were not validated by medical records. Although self reporting is a valid tool to collect accurate data about chronic diseases such as cardiovascular diseases [34, 35] self-reporting could introduce misclassification bias. Had misclassification occurred, it would probably be non-differential with respect to MetS, and thus would reduce the strength of association between MetS and QTc interval duration [21]. More than 50% of the participants who underwent ECG examination were further excluded from the current analysis. They were excluded because they had heart disease or thyroid illness, they were on QTc prolonging drugs or they did not fast for 8 h or more. Persons with cardiac or thyroid disease were not eligible to be included in the analysis because heart disease and thyroid illness are associated with QTc interval prolongation [21]. Fasting is important to detect high triglycerides levels and impaired fasting blood glucose concentration accurately so that individuals who have one or two of these metabolic syndrome components could be captured without misclassification. Hence, participants who did not fast for 8 h or more were excluded. The excluded participants had higher prevalence of heart disease and thyroid illness as well as higher mean systolic arterial blood pressure (P values <0.05). We believe that including such participants would overestimate the magnitude of the association between MetS and QTc interval because the excluded participants are sicker than those who fulfilled the selection criteria.

The cross-sectional nature of these analyses does not allow for inference of causality or for establishment of the temporality between the MetS and QTc interval duration. The QTc interval may be affected by some unknown or known factors not accounted for in this study. Liver cirrhosis and electrolyte disturbance such as hypomagnesemia have been shown to be risk factors for QTc interval [36, 37]. NHANES III did not collect information about these factors. The effect of including such factors in the multivariable analysis is unknown.

Longitudinal studies which collect information about other important QTc prolonging factors are needed to confirm the findings in this study. Additionally, more epidemiological studies are needed to determine whether successful treatment of blood glucose, blood pressure, lipid profile components and waist circumference among persons with metabolic syndrome reduces QTc prolongation and ultimately reduces the risk of ventricular arrhythmia and sudden death in such persons.

In conclusion, this study demonstrates that MetS is an independent risk factor for having a longer QT interval. This study calls for careful ECG monitoring of persons with MetS to enable early detection of QTc interval prolongation for the prevention of severe, and often fatal arrhythmias.

Supplementary material

10654_2008_9252_MOESM1_ESM.doc (26 kb)
(DOC 26 kb)
10654_2008_9252_MOESM2_ESM.doc (27 kb)
(DOC 27 kb)

Copyright information

© Springer Science+Business Media B.V. 2008

Authors and Affiliations

  • Mohammed F. Faramawi
    • 1
  • Rachel P. Wildman
    • 2
  • Jeanette Gustat
    • 3
  • Janet Rice
    • 4
  • Mohammed Y. Abdul Kareem
    • 5
  1. 1.Department of Preventive medicineMenufiya UniversityMenufiyaEgypt
  2. 2.Department of Epidemiology and Population HealthAlbert Einstein College of MedicineNew YorkUSA
  3. 3.Department of EpidemiologyTulane School of Public Health & Tropical Medicine (SPHTM)New OrleansUSA
  4. 4.Department of BiostatisticsTulane School of Public Health & Tropical Medicine (SPHTM)New OrleansUSA
  5. 5.Department of Internal MedicineMenufiya UniversityMenufiyaEgypt

Personalised recommendations