Plasma Somatostatin in Advanced Heart Failure: Association with Cardiac Filling Pressures and Outcome

Chronic heart failure (CHF) is a syndrome involving several organ systems, including the gastrointestinal (GI) tract. In contrast to other heart-organ interactions in heart failure (HF), such as the cardiorenal syndrome, the interaction between cardiac dysfunction and the GI system remains poorly understood. HF patients, however, often present with a broad spectrum of symptoms from the GI tract including abdominal discomfort, anorexia, nausea, early satiety and weight loss. This may result in cardiac cachexia which is a strong independent risk factor for mortality [1]. Theories behind the GI-heart relation in HF have been proposed, including a complex interplay between reduced intestinal blood flow, gut oedema, systemic inflammation, and hormonal imbalance [2], but few studies have examined these theories in patients. In particular, the importance of abnormal hemodynamics, the hallmark of advanced HF, on the heart-gut interaction including hormonal signalling has not been investigated.

Somatostatin is a neuroendocrine peptide widely expressed in the brain and the periphery with a broad spectrum of varying biological effects. In the GI system, the highest amount of the peptide is expressed in the stomach, duodenum, jejunum and pancreas [3]. Locally, somatostatin regulates the GI system by inhibiting the release of most GI- and pancreatic hormones (e.g., gastrin, secretin, cholecystokinin [CCK], neurotensin, glucagon and insulin). Used as an exogenous agent, the somatostatin-analogue octreotide has potent vasoconstrictive effects of the splanchnic vascular bed and is used for instance in patients with portal hypertension due to liver cirrhosis [4].

Given the marked vascular effects of octreotide, we hypothesized that endogenous somatostatin concentrations in plasma would be related to the degree of hemodynamic derangement in patients with advanced HF and, in turn, also with long-term prognosis of these patients.

This was a prospective exploratory study on 53 HF patients with an LVEF <45% scheduled to undergo right heart catheterization (RHC) at the Department of Cardiology at Copenhagen University Hospital, Rigshospitalet. Patients were referred for evaluation for heart transplantation (HTX) or implantation of a left ventricular assist device (LVAD). Since patients were not required to have symptoms of advanced HF at the time of referral, some patients were considered in NYHA II classification at the time of RHC. However, these patients were included, as they had recently experienced advanced HF symptoms and/or their HF condition was considered serious enough to justify referral for further investigation including an invasive cardiac catheterization. Patients were required to be on optimal medical therapy (as tolerated) before referral. For this study, patients were excluded if they have had an AMI within 30 days from the procedure date, if they had active infection, received treatment with an intravenous inotrope at the time of the RHC or if they were younger than 18 years of age. All patients were followed as outpatients. Clinical assessment for signs of congestion were performed on the days of RHC.

The invasive protocol followed routine diagnostic procedure without modifications. Four different experienced physicians performed all the RHCs in the cardiac catheterization laboratory after appropriate zeroing and calibration of the pressure transducer using a Swan-Ganz catheter. The catheter was inserted in the right internal jugular or the right femoral vein where the correct placement of the Swan-Ganz catheter was evaluated by fluoroscopy and by visualization of pressure curves on a monitor.

Patients underwent RHC with determination of cardiac output (CO), cardiac index (CI), central venous pressure (CVP), mean arterial pressure (MAP), mean pulmonary artery pressure (MPAP) and capillary wedge pressures (PCWP). CO was measured by the thermodilution technique. BSA was estimated using DuBois method. CI was determined as cardiac output divided by the body surface area (BSA). MAP was estimated using the formula ([2×diastolic pressure(DBP)]+systolic pressure(SBP))/3 and systemic vascular resistance index was calculated using the formula SVRi = ([MAP-CVP]/CI)×80.

Blood samples were collected after an overnight fasting. Patients were instructed to fast from midnight and blood samples were collected between 7 and 8 in the morning after 30 min of supine rest on the days of RHC.

EDTA-aprotinin tubes were used to collect blood for somatostatin measurements and put on ice immediately after blood withdrawal. Lithium heparin tubes were used to collect blood for NT-proBNP measurements. Somatostatin was measured with an in-house radioimmunoassay using antiserum R-37, which is directed against amino acid sequence 3–7 in somatostatin-14 [5]. Radiolabelled somatostatin-14 was used as tracer, and calibration was performed with synthetic somatostatin-14. Plasma was first extracted using ethanol (70%) followed by decantation and drying the supernatant. The dried supernatant was then reconstituted in assay (barbital) buffer. Given the epitope, the assay measures both somatostatin-14 and somatostatin-28. The working range of the assay is 5–100 pmol/L; CV is <20% across the working range. The upper reference limit in our laboratory is <18 pmol/L, values above this are considered abnormal.

Categorical variables are reported as numbers (n) and percentages (%) and continuous variables are reported as mean ± SD unless indicated otherwise. Continuous data were compared using Student t test, whereas χ² analysis was used for group comparisons of categorical data. Somatostatin and NT-proBNP were log transformed for analyses as they were non-normally distributed. We constructed univariable linear regression analyses to examine the associations between log(somatostatin) and hemodynamic parameters (CI, CVP, MAP, PCWP, MPAP, and SVRi), log(NT-proBNP), eGFR and TAPSE. Multivariable models were constructed with CVP, CI and relevant clinical variables (age, gender, LVEF, BMI, eGFR and log(NT-proBNP)).

End of follow-up date was set to September 1, 2019 and the following events were considered: LVAD implantation, HTX or death (all-cause mortality). Patients could undergo implantation of a LVAD as bridge to transplantation or destination therapy. Kaplan-Meier survival curves were plotted dichomotizing the patients at below and above the normal somatostatin-value of the laboratory (≥18 pmol/L). Furthermore, Kaplan-Meier analysis was performed after dividing patients into tertiles of baseline somatostatin levels. Log-rank test was performed to compare survival distributions between the groups. Cox proportional hazards models were used to estimate hazard ratios to identify predictors of death (censoring patients at time of LVAD or HTX) or predictors of the combined endpoint of death, LVAD implantation or HTX.

ROC curves for somatostatin and prediction of CVP >12 and CI <2.2, respectively, were created and AUC calculated.

Two-sided p values were used, and a p value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS (version 25, IBM Corp.).

Baseline characteristics are presented in Table 1. The study population consisted of 53 patients with a predominance of men (81%). Mean age was 54 years, most patients were overweight (mean BMI 27 ± 5 kg/m²). Mean LVEF was 18 ± 8% and the majority of the patients were in NYHA-class III–IV (79%) at the time of examination. A little more than one-third (36%) of patients presented with at least one clinical sign of congestion such as peripheral oedema or jugular vein distention (JVD). Most patients were treated with recommended HF medications and only a few (15%) were treated with PPIs.

Hemodynamic evaluation confirmed that, on average, filling pressures were elevated but CI was within normal range. NT-proBNP was highly elevated with a median concentration of 290 pmol/L (equivalent to 2.436 pg/mL).

Median somatostatin concentration was 18 pmol/L (IQR 14). An increased somatostatin concentration (defined in our laboratory as ≥18 pmol/L) was present in 27 (51%) patients. Patients with increased somatostatin more frequently had clinical signs of HF, higher NT-proBNP, and a higher frequency of atrial fibrillation. There were no significant differences in age, gender, LVEF, NYHA class, use of PPIs or hemodynamic measurements between the groups.

Univariable and multivariable linear regression models are shown in Table 2. In univariable regression analysis, log(somatostatin) was associated with CVP (r² = 0.14, p = 0.003; shown in Fig. 1) and CI (r² = 0.15, p = 0.004; shown in Fig. 2). Somatostatin was not correlated to pulmonary pressures or SVRi.

Log(somatostatin) remained associated with CVP (p = 0.044) when constructing a “biological” multivariable model adjusting for important clinical variables, that is, age, gender, BMI, eGFR and LVEF; however, when our model was further adjusted for log(NT-proBNP) the association was no longer statistically significant. Somatostatin’s ability to predict abnormal CVP (cutoff value >12 mm Hg) and CI (cutoff value <2.2 mm Hg) is presented in ROC curves (Fig. 3a, b). AUC for the prediction of elevated CVP was 0.775 and for low CI 0.778.

Mean follow-up time was 6.1 years and no patients were lost to follow-up. At the end of follow-up, 17 patients (32%) had died. An LVAD was implanted in 12 (23%) and 15 (28%) were transplanted. Three out of 12 patients who had an LVAD implantation patients were later transplanted. While 21 (40%) were alive with a LVAD or transplant at follow-up, 15 (28%) were alive without.

The results of the univariate and multivariate Cox regression models are presented in Table 3. In univariate analysis, patients with somatostatin levels of 18 pmol/L or more had a 1.6-fold increased risk of death, LVAD or transplantation and a 5.5-fold increased risk of all-cause mortality compared with patients within the reference interval (i.e., <18 pmol/L). In multivariate Cox models adjusted for CVP, CI, age, gender, and NT-proBNP, increased concentrations of somatostatin were not a significant predictor of the combined endpoint. In contrast, multivariate Cox analysis adjusted for the variables mentioned above, identified increased somatostatin concentrations as an independent predictor of mortality with a 4.2-fold increased risk. Kaplan-Meier survival curves are presented in Figure 4a, b.

This is the first study to demonstrate that plasma concentrations of somatostatin, an inhibitory regulator of the GI system, correlate with CVP and CI in patients with advanced heart failure. Further, increased somatostatin concentrations predict all-cause mortality even after adjustment for important clinical variables.

Somatostatin is a regulatory peptide expressed throughout the body where it mediates its actions through 5 receptor subtypes (somatostatin receptor_1–5). Somatostatin is abundantly present in the GI system, and octreotide, a somatostatin analogue used pharmacologically with high affinity for somatostatin receptor₂ and somatostatin receptor₅ [4], is a potent splanchnic vasoconstrictor, decreasing splanchnic blood flow and splanchnic vascular conductance [6]. Octreotides effect is exploited in advanced HF patients treated with an LVAD who have recurrent GI bleeding [7‒9]. Currently, the exact mechanism of octreotide on splanchnic vasoconstriction is yet to be fully understood. Given the vascular effects of exogenous somatostatin, we sought to investigate the relation between endogenous somatostatin and the degree of hemodynamic derangement in patients with advanced HF. Our study suggests that somatostatin might play a role in advanced HF, eluding to a possible gut-heart interaction.

Our findings suggest that hypoperfusion of the intestine and gut oedema – as indicated by reduced CI and elevated CVP – is associated with increased somatostatin concentrations; which in turn could inhibit motility and secretion of GI hormones and thereby contribute to the GI symptoms commonly experienced by HF patients. With the present findings in mind, this hypothesis deserves further examination. Somatostatin plays a role in food regulation. Ghrelin is well established to stimulate food intake where the target for ghrelin orexigenic actions is in the brain. Somatostatins have divergent effects on ghrelin release, in the brain, somatostatin stimulates the increase in ghrelin, whereas in the peripheral, somatostatin has the opposite effect [10]. Further, in animal studies, activation of central somatostatin neurons have proven increased high-calorie intake [11]. It is unknown if peripheral elevated levels of somatostatin could have a direct central effect in appetite regulation and would require further exploration.

The splanchnic vascular compartment is a potential contributor to cardiac decompensation [12, 13]. Decreased abdominal vascular capacitance due to passive and active mechanism controlled by the sympathetic nervous system can lead to volume redistribution. Elevated concentrations of somatostatin could increase GI vasoconstriction which could accentuate GI hypoperfusion in advanced HF and aggravate HF; however, we did not find any correlation between somatostatin and SVRi. While this is less suggestive of a causal effect of somatostatin on GI perfusion in HF, it does not rule it out, as SVRi does not necessarily represent the resistance to flow in the GI bed.

The gut-heart endocrine response in CHF may be more complex than previously assumed and it has recently been shown that cardiomyocytes express the “classic” gut hormone precursor, pro-cholecystokinin (proCCK) a now proven predictor of cardiovascular mortality in HF patients [14‒16]. It is unknown whether the heart also expresses somatostatin, and the information about effects of somatostatin on myocytes is limited. Research has identified somatostatin receptor subtypes in fibroblasts and myocytes in human hearts [17] and suggested a beneficial activation of somatostatin receptor₂ ability to inhibit Ca²⁺ mediated signalling [18] with possible cardio-protective effects. Further studies are, however, needed on human HF-hearts to investigate if the disease alters the expression patterns. Whether an elevated concentration of somatostatin is harmful in HF and a contributing factor to the GI symptoms, or if somatostatin has cardio-protective properties, but is insufficiently elevated in response to cardiac failure (in a manner similar to that of natriuretic peptides), is currently unknown and, indeed, further studies are required to assess the exact function of somatostatin in HF. The association between elevated somatostatin and reduced survival, in a concentration dependent manner, observed in the present study, should inspire future mechanistic studies.

Despite significant improvements in HF management, prediction of outcome remains challenging and the overall prognosis of HF remains poor with life expectancy comparable with many malignant cancers [19]. In clinical practice natriuretic peptides are far from perfect, and new biomarkers for better prediction of outcome as well as new potential targets of therapeutic intervention are urgently needed. The current study suggests that somatostatin is a candidate biomarker in advanced HF, where a predefined abnormal value identified patients with a highly increased mortality, independent of demographic and hemodynamic factors, as well as renal function, and, interestingly, also independently of NT-proBNP. Potentially, as elevated somatostatin appears to identify patients with high CVP, the marker might detect RV dysfunction in patients with advanced HF, a well described predictor of poor prognosis in HF [20] and linked to cardiac cachexia [21].

Hence, the predictive ability of somatostatin may be mediated through this mechanism, even if the effect was corrected for natriuretic peptides and CVP which would be expected to correlate also to RV dysfunction.

In the present study, we found a significant, however relatively weak, association between somatostatin and the hemodynamic variables CVP and CI. Larger studies are needed to explore the relationship between hemodynamic changes in HF and somatostatin, as are studies to investigate the independent prognostic predictive value of somatostatin to establish if there is an incremental value of somatostatin added to predictive clinical variables and natriuretic peptides.

Although the main purpose of the study was to establish a link between hemodynamics and somatostatin in order to better understand HF pathophysiology, we also assessed the ability of somatostatin to predict elevated CVP or reduced CI to evaluate the clinical usefulness of somatostatin as a marker for abnormal hemodynamics. For both parameters acceptable prediction as judged by AUC was found. However, further studies in larger cohorts are needed to confirm use of somatostatin as a surrogate for invasive hemodynamics in clinical practice.

The study included a selected HF patient population that had an indication for RHC referred for evaluation at a single specialized centre exposing the study to the possibility of selection bias. We investigated a considerably younger HF population than average HF patients which may limit the generalizability to the general HF population. None of the patients were treated with neprilysin inhibitors which may affect splanchnic hemodynamics. Future studies should include patients treated with this drug class.

Plasma concentrations of somatostatin are associated with elevated right-sided filling pressure in patients with advanced HF and the pathophysiological mechanism may be related to congestion and/or hypoperfusion of the intestine. In advanced HF, elevated concentrations of somatostatin are independently associated with increased mortality risk.

Compliance with Ethics Guidelines: the study was approved by the Ethics Committee of the Capital Region and the Danish Data Protection Agency and complies with the Declaration of Helsinki. All patients gave written informed consent.

The authors have no conflicts of interest to declare.

The study was funded by the Research Fund at the University Hospital of Copenhagen, Rigshospitalet and the Beckett Foundation.

L.B. and F.G. conceived the original study design. T.D. was responsible for data extraction, analysis and searching for studies. All authors made a substantial contribution to interpreting the data and writing the manuscript. All authors reviewed and approved of the final manuscript before submission.