sST2 Predicts Short Term Therapy Success in Patients with Therapy Resistant Hypertension after Renal Sympathetic Denervation

Abstract

Background: Renal sympathetic denervation (RSD) has provided promising data in its ability to treat therapy resistant arterial hypertension. The effect of RSD on sST-2, a promising biomarker for risk stratification in cardiovascular diseases, has so far not been systematically studied. Methods: We evaluated serum levels of sST-2 and clinical parameter including left ventricular mass (LVM) in 54 patients with resistant hypertension (RH) undergoing bilateral RSD at baseline as well as at one and/or three months. Results: After RSD, mean office blood pressure showed a significant decrease after one month (p < 0.001). On echocardiography a reduction of LVM was observed at three months (p < 0.01). This was accompanied by a significant decrease of sST-2 levels at three months (sST-2 baseline: 6310.1 ± 3246.0 pg/mL vs. sST-2 three months: 4703.8 ± 1585.9 pg/mL, p = 0.048). Furthermore, baseline sST-2 levels were positively correlated with systolic blood pressure at one month (r = 0.514, p < 0.01) but not three months, indicating a potential predictive value of sST-2 for early intervention success. Conclusion: In patients with RH, RSD is associated with a significant decrease of sST-2 levels after three months, indicating sST-2 to be involved in remodeling processes after RSD. Furthermore, lower sST-2 levels at baseline might be a potential predictor of early intervention success of RSD.

1. Introduction

Resistant hypertension (RH), commonly defined as blood pressure (BP) above a treatment goal despite intake of three antihypertensive drugs including a diuretic, is a major clinical burden, affecting 20 to 30% of patients with arterial hypertension [1]. Higher age and obesity are two of the strongest risk factors for uncontrolled hypertension, potentially leading to an increase in the incidence of RH in industrialized countries over time [1]. Re-establishing BP control represents an important goal to lower cardiovascular risk in RH patients.

Chronic sympathetic overactivation has been identified as a driver of RH [1], with sympathetic nerve activation causing increased sodium reuptake, increased renin release, increased arteriolar vasoconstriction and reduced diuresis of the kidney. Interventional percutaneous renal sympathetic denervation (RSD) has been established as a promising interventional treatment approach [2]. Despite divergent results of the Simplicity HTN-3 trial [3], a series of new high-quality sham-controlled trials with favorable outcome have supported the therapeutic value of this procedure [4,5,6]. Furthermore, previous trials also revealed beneficial effects on cardiac reverse remodeling in RH [7]. Successful RSD was able to reverse cardiac hypertrophy and diastolic dysfunction, indicating that RSD has potential implications for cardiac function in hypertensive heart disease, which is a major health burden in this population.

Novel devices for RSD are providing an operator-independent treatment pattern with multiple electrodes or circumferential ablation effects, resulting in a more predictable outcome of the procedure [4,6]. Furthermore, adequate patient selection has been a very important lesson from previous trials. With respect to the lack of an intra-procedural success marker in RSD, reliable markers of treatment success or clinical benefit would be of great value to identify potential patients.

The novel biomarker soluble suppression of tumorigenicity 2 (sST-2) shows promising results for risk stratification in cardiovascular patients. sST-2 belongs to the IL-1 receptor family. It has an influence on immunologic processes with subsequent cardioprotective effects including prevention of myocardial hypertrophy and fibrosis. This indicates a predictive potential in hypertensive heart disease [8]. ST-2 is a protein with two isoforms—a membrane-bound receptor form (ST-2L) and a soluble form (sST-2) secreted by cardiac myocytes exposed to stretch. Interleukin-33 (IL-33) has been identified as the functional ligand of ST-2L. Their interaction is required for the activation of the cellular cascade of events that protect the myocardium from hypertrophy and fibrosis by opposing the effects of angiotensin-II on myocytes. High levels of sST-2 have a detrimental effect due to its action as a decoy receptor by IL-33 neutralization, thus limiting its availability to bind to ST-2L [9]. Therefore, increased sST-2 levels seem to be a predictive indicator for left ventricular (LV) dysfunction. sST2 levels have been shown to be associated with increased systolic blood pressure and structural cardiac alteration including LV hypertrophy and diastolic dysfunction [10]. Therefore, elevated sST-2 levels may reflect hypertensive heart disease in RH and identify patients at risk who might benefit from more aggressive BP lowering interventions such as RSD. In order to investigate a potential application for sST-2 as a predictor of treatment success in RH patients undergoing RSD, serum levels of sST-2 at baseline and during follow up were correlated with clinical variables including arterial blood pressure levels and left ventricular mass (LVM) calculations obtained by echocardiography. We hypothesize sST-2 to reflect relevant clinical outcomes and predict potential success rates of RSD.

2. Materials and Methods

The study was approved by the local ethics committees and was performed in accordance with the Declaration of Helsinki. Following screening and RSD, clinical follow-ups were scheduled at one and/or three months. All patients provided written informed consent for their participation in the trial and further laboratory testing.

Study subjects: In this study, we included 54 consecutive RH patients with available sST-2 measurements at baseline and/or during follow-up at one and/or three months following RSD. Baseline sST-2 measurement was performed before RSD at the time of hospital admission. As eligible patients were confined to those with RH in accordance with previous RSD studies and current hypertension guidelines, the patient collective enrolled was pre-selected and limited in scope. All eligible patients were older than 18 years and presented with a systolic office BP greater than 160 mmHg despite treatment with three or more antihypertensive drugs. No changes in medication were made for the last three months prior to intervention. A preceding 24 h BP measurement was performed to rule out white coat hypertension. Secondary causes of hypertension, as well as renal insufficiency with a glomerular filtration rate <45 mL/min/1.73 m² were excluded before RSD. Further exclusion criteria were pregnancy, type 1 diabetes mellitus, significant valvular heart diseases, myocardial infarction, or stroke within the last six months. Unfavorable anatomical conditions, such as a renal artery diameter of less than 4 mm, length below 20 mm, and an existing or treated renal artery stenosis, were contraindications for the procedure [7].

BP measurements: Before BP measurements, the adherence of the patients to their antihypertensive medication regimen was checked. Office BP was measured in sitting posture using an automatic oscillometric monitor (Omron HEM-705, Omron Healthcare, Vernon Hills, IL, USA), recording BP and heart rate after 10 min of rest on the brachial artery. Three measurements for 1 min intervals were recorded and averaged for statistical analysis. Throughout the study, measurements were performed on the same arm after initial measurements on both sides had ruled out BP differences of more than 10 mmHg.

Transthoracic echocardiography: Transthoracic echocardiography (TTE) was performed with a Philips IE33 ultrasound system (Philips Inc., Amsterdam, The Netherlands) in accordance with the American Society of Echocardiography [11] to assess LVM (n = 45), which was extrapolated from LV linear dimensions by using the Devereux formula [7].

RSD procedure: In 54 consecutive hypertensive patients, renal angiograms were performed via femoral access to conform anatomic eligibility as previously defined [7]. In the same session, RSD was performed using either the Symplicity Flex catheter (Medtronic Inc., Minneapolis, MN, USA) or the Vessix V2 catheter (Vessix Vascular/Boston Scientific, Marlborough, MA, USA). Up to six ablations at 8 W for two minutes each were performed in both renal arteries. In 46 Patients, the intervention was performed with the Symplicity Flex catheter and in 8 patients with the Vessix/Boston Scientific V2 catheter.

Biomarker analysis: In a subgroup of patients, serum levels of sST-2 were analyzed at baseline, at one month (n = 24) and three months (n = 20). Serum levels of sST-2 were measured by using commercially available enzyme linked immunosorbent assay (ELISA) kits (DuoSet ELISA, DY523B, R&D Systems, Minneapolis, MN, USA).

Statistical analysis: Statistical analysis was performed using SPSS (22.0, SPSS Inc., Chicago, IL, USA). All values were provided as mean ± standard deviation (SD). The Kolmogorov–Smirnov test was used to assess a normal distribution of data in the study population. As all parameters and biomarker concentrations were normally distributed, mean values between groups were compared using Student’s t-test. Correlation analysis was performed using Pearson’s correlation analysis. Correlation of sST-2 levels at baseline was only performed in patients if baseline levels of sST-2 were available and at least one follow-up measurement was performed at one month (n = 24) or three months (n = 20). A p-value of <0.05 was considered statistically significant.

3. Results

A total of 54 patients with established RH were included in this study. Most patients were male (72.2%) and the mean age was 59.0 ± 16.6 years. Consistent with previous studies of RH, the patients’ body mass index was either in the upper range of normal or obese; the average BMI was 30.0 ± 4.5 kg/m². On average, patients were taking 3.5 antihypertensive drugs and all patients maintained their baseline antihypertensive drug regimen during the entire study follow-up period to avoid biomarker bias. At baseline, the overall mean sitting BP was 175.6 ± 23.3 mmHg systolic and 94.3 ± 12.2 mmHg diastolic (

Table 1. Baseline characteristics, comorbidities, baseline BP and antihypertensive regime.

Characteristics	n	% or Mean ± SD
Age (years)	54	59.0 ± 16.6
Height (cm)	54	174.5 ± 10.9
Weight (kg)	54	90.9 ± 17.2
BMI (kg/m2)	54	30.0 ± 4.5
Diabetes	19	35.2%
Hypercholesterolemia	32	59.3%
Serum creatinine (mg/dL)	45	0.95 ± 0.3
Systolic BP (mmHg)	54	175.6 ± 23.3
Diastolic BP (mmHg)	54	94.3 ± 12.2
Male sex	39	72.2%
ACE inhibitor	23	42.5%
AT-1 receptor blocker	32	59.3%
Calcium antagonists	24	44.4%
Diuretics/HCT	44	81.5%
Betablockers	43	79.6%
Alpha blockers	11	20.4%
Aldosterone antagonists	7	13%

).

Compared to the baseline office systolic BP (SBP 175.6 ± 23.3 mmHg), a significant reduction in SBP was observed at one month (SBP 157.7 ± 20 mmHg; p < 0.001) and three months (SBP 157.2 ± 24.1 mmHg; p = 0.01) after RSD (Figure 1). Diastolic blood pressure (DBP) recordings also showed a trend towards reduction (DBP 94.3 ± 12.2 mmHg at baseline versus 89.8 ± 14.3 mmHg at one month, p = 0.254, and 87.9 ± 16.4 mmHg at three months, p = 0.084); however, they did not reach statistical significance. In a multifactorial analysis (BMI, serum creatinine, presence of diabetes mellitus Type 2 and serum LDL), diabetes and serum LDL levels revealed a statistical significance on postprocedural BP reduction.

Consistent with a reduction in systolic BP, a steady decrease in calculated LVM was observed during follow-up. At three months post intervention, this reduction reached the level of statistical significance (LVM 246.4 ± 82.4 g at baseline, 216.7 ± 60.6 g at one month, p = 0.163 versus baseline, 180.9 ± 66.8 g at three months, p = 0.010 versus baseline; Figure 2A). These findings may indicate successful cardiac reverse remodeling.

Since previous studies described elevated sST-2 to be associated with increased systolic BP and cardiac remodeling, we further evaluated sST-2 levels at baseline and during follow-up. High sST-2 levels were observed in the RSD population at baseline. Interestingly, similar to LVM alterations, sST-2 levels decreased after three months (6310.1 ± 3246.0 pg/mL at baseline versus 4703.8 ± 1585.9 pg/mL at three months follow-up; p = 0.048), but not one month post-procedure (6310.1 ± 3246.0 pg/mL at baseline versus 5201.8 ± 2569.3 pg/mL at one month follow-up; p = 0.21; Figure 2B).

In the correlation analysis with patients’ characteristics, sST-2 did not show a correlation with age (r = 0.318, p = 0.067), BMI (r = −0.034, p = 0.891), creatinine levels (r = −0.097, 0.643), diabetes (r = 0.328, p = 0.058) and BNP levels (r = 0.118, p = 0.574). Only LDL levels significantly correlated with sST-2 (r = 0.389; p = 0.023).

Therefore, we investigated sST-2 as a potential marker of cardiac remodeling in order to identify patients with a lower chance of BP decrease following the procedure of RSD. Consequently, we correlated baseline sST-2 levels with SBP values during follow-up. We observed a correlation of sST2 levels at baseline with SBP levels at one month (r = 0.514, p < 0.010, Figure 3), but not at three months (r = −0.108, p = 0.652, Figure 4).

4. Discussion

Hypertensive heart disease, characterized by structural alterations like atrial enlargement, left ventricular hypertrophy (LVH) and functional abnormalities like diastolic dysfunction, is a key feature of organ damage in longstanding RH. The presence of LVH is associated with cardiovascular endpoints [12] and an increase in cardiovascular risk independent of in-office or out-of-office BP [13]. Conversely, successful BP reduction in hypertension is associated with a regression of LVH, whereas large meta-analyses have shown this effect to vary significantly between antihypertensive drug regimens [14,15]. Furthermore, if left ventricular reverse remodeling with a reduction in LVM can be achieved in the course of antihypertensive treatment, this is also associated with a reduction in cardiovascular endpoints [12].

With regards to RSD, a multitude of sham controlled clinical trials have been conducted to underscore the potential of RSD to improve blood pressure control in hypertensive patients. These include different technical approaches with second-generation catheters, allowing multisite ablations within the renal arteries using radiofrequency energy [4,5], intravascular ultrasound [6,16] and other techniques [17]. Consistent with these results, our study cohort showed a consistent reduction in mean office systolic BP, reaching the level of statistical significance at three months. According to the 2018 ESC guidelines for management of arterial hypertension [18], only patients with RH had been selected for RSD treatment within the study, which results in a somewhat selective patient cohort. Consistent with previous studies on RSD, higher initial SBP and DBP values were associated with a better therapy response. Similar effects were observed in patients undergoing renal angioplasty [19] and conventional antihypertensive drug treatment [20]. In most previous or ongoing renal denervation studies, the presence of RH with a systolic blood pressure of 160 mmHg or more despite the administration of three or more antihypertensive drugs including a diuretic was required as an inclusion criteria. Due to the small sample size in our cohort, the trend towards DBP reduction did not reach the level of statistical significance.

Given the current evidence supporting the fundamental role of LVH for cardiovascular risk in patients with RH, expectations were high that RSD might also be able to achieve LVH regression on top of improved BP control. Several clinical trials have detected a relevant reduction of the LVM post RSD using TTE [7] and magnetic resonance imaging [21]. Schirmer et al. reported a reduction in the LVM index at six months, showing that the effect of LVH regression was independent of BP reduction [22]. Due to the heterogeneity of trials, meta-analyses may add more reliable data for a prediction of the potential effect of RSD on LVH [23,24]. Wang et al. analyzed 14 observational studies (11 uncontrolled, 3 controlled) and 2 randomized controlled trials investigating LV structural changes after RSD with different follow-up intervals [24]. The authors found that observational studies using TTE reported an average reduction in the LVM index of −13.88 g/m² (the 95% confidence interval CI ranging from −19.94 to −7.82 g/m²) at 6 months and −16.67 g/m² at 12 months (95% CI −25.38 to −7.97 g/m²). In trials using MRI, the effect was smaller, with a mean LVM index reduction of −5.18g/m² at 6 months and −3.00 g/m² (95% CI −11.38 to 5.38 g/m²) at 12 months. Two randomized controlled trials, however, showed no significant change of LVM following RSD [25,26], emphasizing inconsistencies in available data on LVM reduction following BP reduction. Therefore, our results of LVM reduction after RSD in RH patients are well consistent with previous observational studies, but have to be interpreted with caution considering the possibility of operator bias and a small study sample. The time-course of LVM changes, with significant reductions detectable at three months post RSD, are consistent with other settings such as Transcatheter Aortic Valve Implantation (TAVI) implantations, where abrupt changes of myocardial afterload occur in patients with often severe LVH. In patients post TAVI implantations, reductions in LVM have been documented three to six months after valve implantation [27,28,29]. The delay may indicate the time-course of myocardial reverse remodeling to take effect.

Considering the prognostic importance of BP reduction in patients with RH and the unsolved problem of missing intra-procedural endpoints in RSD, a reliable indicator of RSD treatment success would be of enormous clinical value. In this context, biomarkers could be of additional value at least to indicate treatment success or failure at an early stage. Dörr and colleagues have shown promising data using intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) to predict RSD treatment success, showing that 8–12% of patients in their cohort were non-responders [30]. Generally, the results currently available on biomarkers in RSD show a certain disparity of data. Only a limited number of studies investigated the predictive value of biomarker candidates as potential indicators of a successful RSD procedure. Some studies showed a potential predictive value of soluble Fms-like thyrosinkinase-1 (sFLT-1), ICAM-1, VCAM, while other biomarkers, including Galectin-3, Copeptin, soluble vascular endothelial growth factor receptor 1 (sVEGFR-1), vascular endothelial growth factor -A/-C (VEGF-A/-C), nitric oxide (NO) and soluble vascular adhesion molecule 1 failed to predict BP lowering effects after RSD.

sST-2 is a novel biomarker, which has been proven as very valuable for the prediction of clinical endpoints in acute and chronic cardiac conditions independent of classic cardiac biomarkers as Troponin-T. With regards to patients with high BP and RH, the gradual development from hypertension to LVH and further to hypertensive heart disease with diastolic heart failure corresponds to different serum levels of sST-2. Furthermore, sST-2 is an indicator of cardiac congestion. Ojji et al. showed in a cohort of 210 patients that the serum sST-2 levels differentiated between different stages of hypertensive heart disease [31]. The serum levels of sST-2 in patients with hypertensive heart failure were higher than in patients with hypertensive heart disease without heart failure, sensitivity and specificity, at 76.5% and 100%, respectively. Furthermore, within 133 patients with arterial hypertension, those with LVH had higher sST2 levels than those without LVH (23.8 ± 8.33 ng/mL vs. 14.5 ± 4.9 ng/mL, p < 0.001) [31].

In our cohort, compared to physiological levels described in the literature (the reference range of sST-2 was 5.7–53.5 µg/L for men and 4.4–42.4 µg/L for women) [32,33], high sST-2 levels were observed at baseline. Consistent with observations regarding reverse remodeling of LVH, a steady decrease was observed that reached statistical significance at three months post intervention. Therefore, consistent with data from patients with hypertensive heart failure, the decrease in sST-2 might reflect a cardiac reverse remodeling process and following a reduction of cardiac afterload. This deduction is consistent with similar observations and a comparable timeframe of reverse ventricular remodeling in TAVI patients. Of note, successful intervention and a decrease of cardiac afterload in the TAVI-population was associated with a reduction of sST-2 levels, which was not present in the early phase after the procedure (one week after TAVI) [34].

Based on this observation, we studied the value of ST-2 as a potential indicator of pathologic cardiac remodeling to identify patients with pronounced pathophysiological cardiac alterations prior to RSD and likely lower success rates. We observed a positive correlation of ST-2 levels at baseline with SBP levels at one month, but not at three months. The most obvious explanation would be the small sample size compromising complex statistics. On the other hand, it has been established that higher sST-2 levels in RH at baseline reflect a pathologically remodeled left ventricle. In the early phase after procedure in these patients, reduced cardiac afterload due to BP reduction may have promoted LV hypercontractility with a subsequent increase in cardiac output. Therefore, despite a successful decrease of sympathetic activation, higher SBP levels were sustained in the early follow-up phase. However, cardiac output might normalize with further progress of a positive remodeling processes after RSD over time. This would also promote lower SBP levels in these patients. Nevertheless, as we cannot confirm our assumption, this hypothesis should be considered with caution. Our deduction, however, is consistent with observations of a somewhat lower initial effect of blood pressure decrease post RSD [4,7,35] and with higher initial blood pressure values in TAVI patients due to LV hypercontractility in the early phase after TAVI implantation, while blood pressure levels show a decrease during later follow-ups [36]. Therefore, sST-2 might be of predictive value for patients undergoing RSD. Of note, as a short-term outcome marker, application of this biomarker might help to identify potential candidates where more rapid BP lowering measures are required. However, since we did not analyze the prognostic value of sST-2 in our population with regards to clinical endpoints, these speculations require further investigation. Larger studies with longer clinical follow-ups are needed to assess the predictive value of sST2 in RSD patients.

5. Conclusions

To our knowledge this is the first study to investigate the correlation of BP and LVM changes with the novel cardiac biomarker sST-2. In patients with RH we could detect a decrease in serum sST-2 levels following interventional RSD, concomitant with BP reduction and LVM regression. These changes reached the level of statistical significance after three months, potentially indicating a reduction in myocardial distension and afterload as a driver of reverse cardiac remodeling.

Furthermore, sST-2 levels at baseline might be a potential predictor of early intervention success of RSD. Further studies are warranted to clarify the role of sST-2 in the treatment of hypertensive patients and to study the value of sST-2 in RSD.

Our study suffers from some limitations, including the small cohort as well as the absence of a control group undergoing a sham procedure. The retrospective character and associated incomplete representation of the entire cohort in the biomarker subgroup are further limitations potentially causing bias of the results. Of 54 patients enrolled in the study, 20 patients completed the entire follow-up investigations, including sST-2 biomarker measurement. In accordance with previous studies in RH, patients with established RH were enrolled and patients with type 1 diabetes mellitus, renal insufficiency with a glomerular filtration rate <45 mL/min/1.73 m² and with secondary causes of hypertension were excluded. Therefore, the patient cohort was highly selected. Nevertheless, this is the first trial exploring the application of sST-2 in the context of RSD in a population with RH. While we present some promising results, larger-scale studies are required to confirm our observations and to further investigate the utility of this biomarker in RSD.