|Home | About | Journals | Submit | Contact Us | Français|
Aims: Non-alcoholic hepatic steatosis (HS) is associated with hypertension and increased cardiovascular risk. While Blood pressure hyper-reactive response (HRR) during peak exercise indicates an increased risk of incident hypertension and increased cardiovascular risk, no data on the association of non-alcoholic HS and HRR exists. In this study, we have evaluated the association of HS with HRR.
Methods: We included 13 410 consecutive individuals with a mean age: 42.4 ±8.9 years, 3561 (26.6%) female with normal resting blood pressure and without a previous diagnosis of hypertension, who underwent symptom limited exercise treadmill test, abdominal ultrasonography and clinical and laboratory evaluation. HS was detected by abdominal ultrasonography. HRR was defined by a peak exercise systolic blood pressure>220mmHg and/or elevation of 15mmHg or more in diastolic blood pressure from rest to peak exercise.
Results: The prevalence of HS was 29.5% (n=3956). Overall, 4.6% (n =619) of the study population presented a HRR. Subjects with HS had a higher prevalence of HRR (8.1 vs. 3.1%, odds ratio 2.8, 95% CI 2.4—3.3, P < 0.001). After adjustment for body mass index, waist circumference, fasting plasma glucose and low density lipoprotein cholesterol, HS (odds ratio 1.4, 95% CI 1.1–1.6, P =0.002) remained independently associated with HRR. HS was additive to obesity markers in predicting exercise HRR.
Conclusions: Non-alcoholic HS is independently associated with hyper-reactive exercise blood pressure response.
Non-alcoholic fatty liver disease (NAFLD) includes a spectrum of liver abnormalities typically seen in individuals with obesity, metabolic syndrome, insulin resistance and diabetes.1 Although some of the dysmetabolic changes detected in this scenario may also occur in alcoholic fatty liver disease, e. g. hypertriglyceridemia, other characteristics of the atherogenic dyslipidemia are more prevalent in individuals with NAFLD.2 HS one of the components of NAFLD, is strongly associated with other makers of increased cardiovascular risk such as systemic arterial hypertension,3 insulin resistance,4 low-grade inflammation,5 atherogenic dyslipidemia6 and metabolic syndrome.7 Mechanistic pathways underlying this association, however, are poorly understood. It is not definitively established if HS and hypertension constitute mere co-phenomena in the context of insulin resistance or if HS may actually be directly involved in the pathogenesis of hypertension.8 Indeed, it has been suggested that active secretion of cytokines by the fatty liver can interfere with blood pressure homeostasis through modulation of sympathetic activity, endothelial dysfunction and arterial stiffness.9,10
Despite the association of HS with most of the components of the metabolic syndrome, the evidence of its association with hypertension is less well documented.11 Nonetheless, one meta-analysis has demonstrated that hypertension is associated with the future development of hepatic fibrosis in individuals with NAFLD at baseline.12 Additionally, one large study has demonstrated that the presence of NAFLD at baseline is associated with the development of incident overt hypertension.13 However, no study to date has evaluated whether this association is also present for individuals in conditions associated with future development of hypertension.
Prospective studies have shown that hypertensive response during peak exercise in otherwise normotensive individuals is associated with an increased risk of future development of hypertension.14 Additionally, prior evidence suggests that hyper reactive blood pressure response (HRR) during the exercise treadmill test is associated with increased risk of cardiovascular events.15–17 However, the association of HRR as a marker of abnormal blood pressure response, with NAFLD or HS has not been studied. In the present study, we have evaluated whether HS is independently associated with HRR during exercise. We have hypothesized that blood pressure abnormalities such as HRR, are independently associated with the presence of HS even in individuals with normal blood pressure.
We prospectively included all consecutive individuals with no prior history of cardiovascular disease, overt hepatic disease or hypertension who participated in a voluntary comprehensive health evaluation paid by their employers. We included all individuals who underwent symptom limited exercise treadmill test, abdominal ultrasonography and an extensive clinical and laboratory evaluation as part of a routine executive health assessment protocol between 2006 and 2012 at the Preventive Medicine Center of the Hospital Israelita Albert Einstein in Sao Paulo, Brazil. Individuals with incomplete data regarding the covariates of interest, those who presented moderate/high alcohol intake, and those with a resting blood pressure≥140/90mmHg or under antihypertensive drug therapy at the time of the evaluation were excluded (N=3016). The study protocol was approved by the local Institutional Review Board and was granted a waiver of informed consent.
The clinical, laboratory and health test procedures of the check-up protocol were previously described.5 In short, demographics, medical history and medication use were routinely recorded by a standard questionnaire. Alcohol intake was assessed by the Alcohol Use Disorders Identification Test (AUDIT) questionnaire and alcohol abuse was defined by AUDIT score≥8.18 Since excessive alcohol consumption is a common etiology of both HS and hypertension and could explain the association of these conditions,19 those with an AUDIT score≥8 were systematically excluded from the study.
Smoking status was categorized as current smoker (at least 1 cigarette during the last 30 days) versus current non-smoker. Physical activity was assessed by the International Physical Activity Questionnaire (IPAQ) and subjects were classified as sedentary, poorly active, active and very active.20
Blood pressure was measured three times at the sitting position with an aneroid sphygmomanometer according to the standard method recommended by the American Heart Association.21 Hypertension was defined by average blood pressure values ≥140/90mmHg during the evaluation or by the use of blood pressure lowering medications. Pre-hypertension was defined by a systolic pressure from 120 to 139mmHg and/or a diastolic pressure from 80 to 89mmHg.
Height (m) and weight (kg) were measured with a standard physician's scale and a stadiometer in order to calculate body mass index (BMI, kg/m2). Waist circumference was recorded at the smallest diameter between the iliac crest and the costal margin with a plastic anthropometric tape held parallel to the ground. Obesity and overweight were grouped together and defined by a BMI >2 5kg/m2, while abdominal obesity was characterized by elevated waist circumference (>80cm in women and >94cm in men). The metabolic syndrome was defined according to joint IDF/AHA criteria,22 by the presence of ≥3 of the following elements: elevated waist circumference (WC); elevated blood pressure (systolic ≥130mmHg and/or diastolic ≥85mmHg, or use of antihypertensive medications); fasting blood glucose level ≥5.55mmol/l or use of drug treatment for hyperglycemia; elevated triglycerides level (≥1.7mmol/l); and low HDL-C level (<1.02mmol/l for men or<1.28mmol/l for women) or the use of lipid-lowering medications.
Blood samples were collected after a 12-h fasting and processed at the Central Laboratory of the Preventive Medicine Unit of the Hospital Israelita Albert Einstein. Total cholesterol, triglycerides (TG), HDL cholesterol, glucose and liver transaminases were determined using standardized automated laboratory tests (Vitros 5600, Johnson & Johnson Orthoclinical Diagnostics). When TG<4.5mmol/l, low density lipoprotein (LDL) cholesterol was calculated by the Friedwald formula. When TG≥4.5mmol/l, LDL-cholesterol levels were measured directly. High-sensitivity C reactive protein (hsCRP) levels were assessed by immunonephelometry (Dade-Behring). Estimated 10-year risk of hard coronary heart disease events was calculated by the original Framingham risk equations.23
All participants underwent regular exercise treadmill test using either Bruce or Ellestad Protocols, according to the individual physical conditioning. Subjects remained on the treadmill for up to five 3-min stages. They exercised until reaching an age-specific target heart rate or the development of symptoms requiring termination of the test.24 Systolic and diastolic blood pressures were recorded by cuff when the subject was standing immediately before testing and during the last minute of each 3-min exercise stage. HRR was defined by a peak exercise systolic blood pressure >220mmHg and/or elevation ≥15mmHg in diastolic blood pressure from rest to peak exercise, corresponding to sex-specific, age-predicted BP ≥95th percentile.14
All subjects underwent abdominal ultrasonography using a Siemens ACUSON XP-10 device (Mountain View, CA) after a minimum 6-h fasting in order to screen for HS. HS was identified by board certified radiologists according to conventional ultrasonographic criteria,25,26 focusing on a pattern of a bright liver, with evident contrast between hepatic and renal parenchyma. A senior radiologist reviewed all the ultrasounds.
Differences in baseline characteristics of individuals with and without HRR on the exercise test were evaluated using independent sample t test or rank-sum test for continuous variables and chi-square test for categorical variables. Continuous variables were described by mean and standard deviation or median and quartiles, as appropriate. Categorical variables were presented as absolute numbers and proportions. The association between HRR and non-alcoholic HS was tested in a multivariate analysis initially adjusted for age and gender with logistic regression models. Covariates included age, gender, smoking status, BMI, waist circumference, physical activity, metabolic syndrome and the presence of prehypertension. Subsequently, all the relevant variables, including HS, were grouped and tested in a multinomial multiple regression model. Tests were conducted considering a significance of 5%. All analysis were performed with SPSS version 20.0.
The final sample consisted of 13 410 subjects with a mean age of 42.4±8.9 years, and 3561 (26.6%) female. Table 1 shows the demographic, clinical and laboratory characteristic of the study population, which consisted of mostly young adults with a low burden of cardiovascular risk factors and a consequent low cardiovascular risk profile, as estimated by the Framingham Risk Score. Among them, only 1040 (7.8%) were current smokers and only 234 (1.7%) were diabetics. Prehypertension was highly prevalent (56.9%), reaching 76% among those with HS. The overall prevalence of metabolic syndrome and of HS were, respectively, 17.2% (n=3211) and 29.5% (n = 3956).
As expected, those with HS were older; more frequently male; and presented higher BMI, WC and resting blood pressure values. Individuals with HS showed higher prevalence of prehypertension, diabetes, smoking and metabolic syndrome. In addition, HS subjects presented higher values of total cholesterol, triglycerides, fasting glucose and hsCRP, while lower values of HDL-cholesterol levels. As a consequence, the prevalence of subjects with an estimated 10-year risk of coronary events>10% was higher in the HS group.
Overall, 4.6% (n=619) of the study population presented HRR on the exercise test. HRR was associated with age, gender and all clinical and laboratory risk factors for cardiovascular disease, except current smoking (P=0.67) (Table 2).
Subjects with HS had a higher prevalence of HRR than those without it (8.1 vs. 3.1%, odds ratio OR 2.77, 95% confidence interval CI 2.36 to 3.26, P<0.001) on the univariable analysis. Other variables associated with HRR were BMI>25kg/m2 (OR 3.08, 95% CI 2.54–3.75, P<0.001); fasting glucose>5.55mmol/l (OR 2.29, CI 95% 1.82–2.89, P<0.001); elevated waist circumference (OR 2.78, 95% CI 2.33 to 3.32, P = 0.001); and low HDL-cholesterol levels (OR 1.28, 95% CI 1.08–1.52, P =0.005)].
Table 3 shows predictors of HRR after adjustment for age and gender, while Table 4 shows the fully adjusted model. After adjustment for these variables, HS remained independently associated with HRR (OR 1.35, 95% CI 1.12–1.62, P = 0.002). Because HS and overweight/obesity, defined by BMI>25kg/m2 or by the presence of an elevated waist circumference, were independent predictors of exercise-induced HRR, their additive effect was tested. While BMI and non-alcoholic HS are associated with a 2.5–3-fold increase in HRR when alone, their joint association resulted in a 4-fold higher prevalence (Figure 1). Findings of similar magnitude were encountered when the association of HS and increased waist circumference were tested on the prevalence of HRR (Figure 2).
In this large sample of normotensive and apparently healthy individuals, HS was strongly and independently associated with blood pressure hyper reactivity on the exercise treadmill test. To the best of our knowledge, this is the first report to disclose the association between HS and an early pre-hypertensive stage such as exercise-induced hypertension.
In a previous study, Huot et al.27 evaluated 317 apparently healthy individuals and found that the relationships among insulin resistance, low cardiorespiratory fitness and increased exercise blood pressure was mediated by an elevated waist circumference. Our findings extend that relationship to the fatty liver as a peculiar kind of ectopic fat depot. Interestingly, the role of fatty liver in HRR prediction was independent and additive to waist circumference or BMI, suggesting that each one of them may directly contribute to hypertension onset, eventually overlapping and boosting each other effect in a synergistic interaction. This may also be valid, in a lesser degree, to other minor kinds of ectopic fat depots, such as intra-thoracic adipose tissue.28
The direct association between NAFLD and hypertension has been studied previously. Prior studies have demonstrated that NAFLD is associated with future development of hypertension, particularly for individuals with moderate or severe NAFLD.13,29 On the other hand, the presence of hypertension is associated with increased risk of progression from NAFLD to fibrosis.12 Our current results build on this prior knowledge by demonstrating that this association between blood pressure and HS extends beyond the range defined as hypertension, suggesting a continuous association between stages of the hypertensive disease and NAFLD development, progression and complication.
Although HS is one of the most common causes of end-stage liver disease, individuals with HS are more likely to die from cardiovascular disease rather than from liver-related causes.30,31 HS carriers are at higher cardiovascular risk and the increased mortality of this group results primarily from cardiovascular disease.32 Given that HS can be identified in about one adult in four, the excess risk imposed by this condition represents a concerning public health problem. However, whether HS is just a marker of cardiometabolic risk or it plays a direct role in atherogenesis and atherothrombotic risk is still matter of research.
Several studies already demonstrated that the presence of HS might predict clinical33 and subclinical atherosclerosis,34 even after adjustment for behavioral risk factors, all of the components of the metabolic syndrome and the presence of central obesity. In this context, our results points HS interference in blood pressure homeostasis as one of the possible explanations for this excess risk.
Several mechanisms may explain the association of increased blood pressure response and HS demonstrated in our study. Previous studies suggest sympathetic activity may be one of those pathophysiological links, as increased sympathetic activity has been demonstrated to be associated with the HRR, and the future development of hypertension, while it is also linked to insulin resistance.35–37 Additionally, both HRR and HS are associated with endothelial dysfunction,31,38,39 though it is not yet clear if this is simply an association due to clustering of risk factors in those individuals or if this may be a pathophysiological link between the two.
Despite the robust results, our study presents some limitations. Its cross-sectional design does not allow a prospective evaluation of cardiovascular events and hypertension onset. However, previous longitudinal studies already addressed such end-points among individuals with HRR. HS evaluation by ultrasonography is an issue of concern for many hepatologists due to its possibly limited accuracy in diagnosing fatty liver.40 However, its use has been previously validated and a high degree of concordance with magnetic resonance and liver biopsy was demonstrated.41 Furthermore, abdominal ultrasonography is widely available, inexpensive and non-invasive, which make it an attractive screening tool in real life clinical practice. Additionally, despite adjustment for self-reported physical activity, metabolic equivalents during the treadmill test would have been a more reliable measurement of the cardiorespiratory fitness of our study population.42 However, since it was previously demonstrated that visceral adiposity modulates blood pressure response irrespectively of cardiorespiratory fitness,27 it is unlikely that such data would change our main findings. Additionally, ambulatory blood pressure measurements a better tool for the diagnosis of masked hypertension were not performed due to the nature of this study and size of the population. Not only that, the present cohort may have included a small number of individuals with chronic viral hepatitis, as this was not routinely ruled out in our cohort. Nevertheless, the prevalence of viral infection in our center is below 1%. Similarly, there is a potential for misclassification of some individuals who may have alcoholic HS, as we have relied on self-reported alcohol consumption for the AUDIT score. Nonetheless, the risk for misclassification based on the AUDIT score is low. Thus, this potential lack of precision in the definition of the etiology is unlikely to have significantly influenced the results. Finally, given that our study population included young Brazilian adults, our findings may not be fully generalizable across other ethnicities or populations with broader age ranges. Nevertheless, HS prevalence in our population is consistent with previous reports from other countries43 pointing similar epidemiologic profiles.
In conclusion, HS was independently associated with hyper reactive blood pressure response on the exercise treadmill test. This association may partially explain the increased cardiovascular risk observed in HS; enlighten the pathophysiological link between fatty liver and hypertension; and underscore the need for hypertension screening among HS participants.
Conflict of interest: R.D.S. has consulted and received honoraria (modest) from: Astra Zeneca, Amgen, Akcea, Aegerion, Biolab, Boehringer Ingelheim, Eli-Lilly, Genzyme, Kowa, Merck, Sanofi/Regeneron, Pfizer, Praxis, Torrent and Unilever; M.J.B. has consulted for Pfizer and Akcea (modest honoraria).