The complex interplay between obesity and hypertension has been a matter of interest and debate for a long time. In the last decade several new insights into the pathophysiology linking obesity and hypertension have led to a better understanding of the mechanisms that beneath this association: insulin resistance, activation of the renin-angiotensin-aldosterone system (RAAS), dysfunctional adipose tissue and deregulated synthesis of adipokines, increased activity of the sympathetic nervous system, overproduction of inflammatory cytokines and obstructive sleep apnea are among some of these factors [1]. Together, these neuro-endocrine imbalances contribute to vascular and endothelial dysfunction, impaired pressure natriuresis and sodium excretion, increased cardiac output and changes in systemic vascular resistance and arterial compliance [2]. In this respect, the measurement of hemodynamic parameters may be a valuable tool for the management of obese hypertensive patients, not only to identify the main abnormalities associated with elevated blood pressure (increased CI, SVRI or fluid volume) [3] but also to help guide therapeutic adjustments according to the spectrum of hemodynamic changes identified [4].

It is also worthwhile to note that obesity is one of the most frequent causes of resistant hypertension, and identification of major hemodynamic abnormalities may help shorten the time to achieve blood pressure control and reduce global cardiovascular risk [3].

While invasive procedures are not adequate for ambulatory patients, non-invasive monitoring may be easily implemented using accurate and reproducible techniques like impedance cardiography [5]. Previous studies have shown the value of this technique in tailoring hypertension treatment in individual patients [6] and several trials have used ICG in large cohorts of hypertensive’s to identify factors associated with resistant hypertension [7] or to better achieve blood pressure control in patients with uncontrolled hypertension [3, 8]. We aimed to identify hemodynamic variables associated with poor blood pressure control in a population of obese hypertensive patients using ICG and to categorize the main changes in hemodynamic parameters associated with high BMI.

One hundred patients were male (49.5%) and average age 54.6 ± 13.9 years. Average systolic and diastolic pressures were 136.5 ± 22.4 mmHg and 82.9 ± 5.1 mmHg, respectively. The average BMI was 28.9 ± 5.1 Kg/m²: 46 patients (22.7%) had BMI between 20 and 24.9 Kg/m², 78 (38.6%) between 25 and 29.9 Kg/m² and in 75 patients (37.1%) BMI was ≥30 Kg/m². According to gender the average values of BMI were 28.5 ± 4.5 Kg/m² for men and 29.5 ± 5.4 Kg/m² for women (p < 0.001). Seventy one patients (35.1%) had systolic AP ≥140 mmHg and 45 patients (22.3%) diastolic AP ≥90 mmHg. BMI correlated with diastolic AP (Pearson’s coefficient 0.163; p < 0.001) but the association was stronger for systolic AP (Pearson’s coefficient 0.235; p < 0.001). There was also a positive correlation between obesity and uncontrolled AP (Pearson’s coefficient 0.138; p < 0.001), although this correlation was weak.

No adverse effects secondary to the ICG procedure were reported.

Regarding SACI, the odds associated with uncontrolled HT was − 1.383, pointing to reduced compliance as a predictor of difficult to control AP. Besides, SACI was reduced not only in obese patients with uncontrolled HT but also with increasing weight.

Obese hypertensives present increased values of CI and SVRI in comparison with non-obese patients. However, the most important hemodynamic determinant of uncontrolled systolic and/or diastolic AP is increased CI. Taking into account that AP is the product of CO and SVR, and using weight-adjusted variables, our results show that in the obese hypertensive patient variations in CI are more important in determining the level of AP.

In obese patients SAP and DAP were greater than in non-obese, reflecting that in these patient population there is increased risk of uncontrolled HT for which seem to contribute both CI and SVRI. There was also increased LCWI in the obese patients. However, this variable, which results from the product of mean arterial pressure and CI, reflects more the higher oxygen demands of the heart in these circumstances than increased cardiac performance itself. Probably, this is also the reason why, in multivariate analysis, LCWI was not predictive of uncontrolled HT as it relates more with coronary ischemic threshold and does not represent a true measure of contractility.

In this way, changes in CI seem to be the most appropriate to guide therapeutic adjustments in the hypertensive obese population studied. These results are in accordance with others previously published [13–16], where obesity is frequently associated with increased cardiac output and CI to accomplish with the higher metabolic demands. In this context, the choice of anti-hypertensive medication should focus more on drugs like beta-blockers or central acting agents [3, 7] that may decrease cardiac output and the effort performed by the myocardium to pump blood through the vessels. Concerns about weight gain related to beta-blockers should be considered, since several trials have documented an association between beta-blockers, weight gain and accumulation of visceral adiposity [17, 18]. Central agonists, to a lesser extent, may be a second choice mainly due to a more unfavorable profile of side effects.

The SVRI values, with a normal range between 1970 and 2390 dyne.sec.cm^− 5.m², were systematically high across all groups of patients (normal weight, overweight and obesity). The differences were significant between non-obese and obese and became even more pronounced when the cut-off ≥140/90 mmHg was established. In this regard, decreasing SVRI, which is an indirect measure of cardiac afterload, should be part of the anti-hypertensive regimen using agents like blockers of the renin-angiotensin-aldosterone system (RAAS) and calcium channel blockers.

Remarkably, however, SVRI was not predictive of uncontrolled HT in the obese patient in multivariate analysis. Although it is expected that vasoconstriction contributes to high levels of blood pressure, several reasons may explain why we didn’t observe that in our cohort: the increase in AP levels was not paralleled by the increase in SVRI and the absence of knowledge on current anti-hypertensive medication, specially with drugs that interfere with arterial tonus, does not to allow to interpret the heterogeneity of SVRI values found. However, it is also recognized that a blunted reduction of peripheral resistance, despite the increased stroke volume and CI, is a common finding in obesity [14, 16] that may contribute to the high levels of SVRI observed. Some factors that may explain this blunted reduction in the tonus of resistance vessels include activation of the sympathetic nervous system and release of substances from the adipocytes [1] leading to high levels of arterial resistance.

Considering SACI, we observed lower levels in the obese hypertensive, especially in the patients with uncontrolled AP and, in multivariate analysis, this variable was predictive of uncontrolled AP. Lower levels of SACI, representing an indirect measure arterial stiffness, were associated with greater risk of blood pressure ≥ 140/90 mmHg. Considering that SACI results from the ratio of CI and pulse pressure it is reasonable to expect, especially in those with higher levels of blood pressure, higher values of pulse pressure and, concomitantly, lower levels of SACI.

It is also worthwhile to note that obese patients, in general, and those with uncontrolled blood pressure had lower values of TFC. Although these data do not point to hypervolemia as a determinant factor to high blood pressure in this cohort, the absence of data concerning use of diuretics limits the interpretation of these values. In the model of multivariate analysis we didn’t included this variable, as expressed in Kohm and not adjusted to body surface area, doesn’t reflect volume/m². Nevertheless, variations of TFC may be very sensitive and consistent with volume expansion [19]. The usefulness of TFC measurements have been previously documented in other studies, proving beneficial in hypertensive patients to help guide diuretic adjustments in those who present with increased chest fluid volume and hypervolemia [19–21] and also in established heart failure, to identify patients at risk for acute decompensation [22, 23].

There are limitations to this work since its retrospective nature and several relevant data missing. One of the most important in this context concerns information on current anti-hypertensive medication on the date of the ICG examination, which we were not able to fully collect due to technical problems concerning data archives and lack of data on waist circumference and abdominal obesity. BMI, although reflecting weight excess and being the basis for obesity grading, does not allow to identify patients with predominantly central adiposity. The evidence in the literature regarding the pathogenic role of abdominal fat is vast and the contribution of insulin resistance, oxidative stress and deregulated synthesis of hormones by adipose tissue for cardiovascular disease is well documented [24–28]. In our cohort we observed high prevalence of overweight and obesity and, probably, stratifying patients according to values of waist circumference would allow improvement in identifying the hemodynamic pattern associated with this type of obesity, more frequently associated with HT.

In obese patients, trends in CI and SVRI are clearly distinct from the one’s we found in lean subjects (which presented a predominance of increased SVRI in spite of CI) and it would be important to investigate if central obesity is associated with extreme values of the considered hemodynamic variables or, in alternative, with variations in other parameters that would impose changes in anti-hypertensive therapy. It should also be mentioned that when considering abdominal obesity, in addition to waist circumference, other elements should also be taken into consideration, namely the differential composition of abdominal fat (subcutaneous versus visceral fat) and the proportion of fat-free mass. While some studies support the importance of fat-free mass as a major determinant of the variations in SV and CO to accomplish with greater metabolic demands in the obese patient [14, 29] others suggest the predominant role of fat mass in the disturbance of hemodynamics [13]. Given the intense metabolic activity of adipocytes it is expected that the several molecules they secrete may interfere with systemic hemodynamic, inducing a state of high CI and blunted decrease in SVRI. Future research including a greater number of obese patients, complemented with data of body composition in terms of relative fat distribution, would allow for a better understanding of hemodynamic patterns associated with different types of obesity.

It is also important to refer two other limitations: first, patients with cardiac arrhythmias were excluded from this study and, in the general hypertensive population they represent a large proportion of patients. In this regard, as the study of hemodynamic in this specific subset is unreliable with ICG, non-invasive determination of CI and SVRI is compromised in these patients. Second, when classifying patients with controlled versus uncontrolled HT we based exclusively on the values of AP obtained in the day of examination. Information about number of medications in each group, time of follow-up in the Hypertension Clinic or time to achieve blood pressure control would better describe the differences between the two groups.

Tailored anti-hypertensive therapy for the obese patient should focus in the balance between CI and SVRI in each patient. In our cohort, anti-hypertensive drugs aimed to modulate cardiac index, mainly beta-blockers, and decrease SVRI like inhibitors of the RAAS and calcium channel blockers seem the most appropriate to control blood pressure. Increased levels of CI contribute to higher risk of uncontrolled HT in the obese, situation in which adjustment of drugs that reduce hyperinotropy may improve and shorten time to blood pressure control.