Septic cardiomyopathy: evidence for a reduced force-generating capacity of human atrial myocardium in acute infective endocarditis

Introduction

Infective endocarditis (IE) is an endovascular microbial infection of cardiovascular structures. The characteristic lesions consist of vegetations composed of thrombotic platelets, microorganisms (bacteria), fibrin, and inflammatory cells [1]. Despite improvements in surgical and anti-infective therapy, IE still carries a high morbidity and mortality [2]. When the aortic and/or mitral valves are destroyed by bacterial vegetations, left heart decompensation caused by volume and/or pressure overload may be present. Reflex tachycardia and depressed peripheral resistance may cause additional damage to the myocardium.

Acute IE may be defined in terms of a multifactorial systemic disease that is associated with systemic inflammatory response syndrome (SIRS) secondary to the infection. SIRS is typically accompanied by arterial hypotension, caused by profound systemic vasodilatation, and multiple organ dysfunction [3]. It represents a state of acute circulatory failure characterized by persistent arterial hypotension despite adequate fluid resuscitation or tissue hypoperfusion (as evidenced by lactate levels >4 mg/dL) not explained by other causes [4].

According to Drosatos et al. [5], cardiac dysfunction plays an important role in the development of multiorgan failure caused by severe sepsis. They observed that mortality was higher in septic patients with systolic and diastolic cardiac dysfunction than in septic patients without cardiac dysfunction. The exact pathophysiological pathways underlying septic circulation patterns or septic cardiomyopathy (SCMP), however, remain to be clarified.

To better understand the hemodynamics of a septic patient, the analogy between the heart and a hydraulic pump serving to produce the cardiac output may be helpful. The work such a pump is capable of rendering is determined by the interaction of preload, global ventricular performance, and afterload [6]. Septic shock, on the one hand, is associated with an increased heart rate and a reduced afterload resulting in an increase in cardiac output. On the other hand, there is relative hypovolemia reducing the preload and the left ventricular end-diastolic volume. Reduced myocardial stretch caused by less end-diastolic volume bates cardiac output [3].

SCMP requires some consideration as well. SCMP may be induced by IE as demonstrated by animal studies showing that the vegetation-associated host (bacteria) protease activity not only leads to proteolytic tissue damage, as evidenced by the release of extracellular matrix fragments such as fibrinogen, but also induces inflammation and myocardial cell damage [7]. In this respect, it is essential to note that SCMP can be caused not only by IE but also by any infective focus in the organism that gives rise to systemic endotoxemia [8].

The mechanisms by which sepsis causes cardiac dysfunction are not completely understood yet. Released in large quantities in patients with septic hemodynamics, circulating inflammatory cytokines such as interleukins or tumor necrosis factor-α (TNF-α) may, for example, be among the factors accounting for SCMP [5]. Clinical studies in humans investigating the pathways of SCMP are therefore subject to considerable limitations due to a variety of in vivo influences exerted not only by intrinsic factors but also by hemodynamic alterations caused by sepsis [3], [6], [9], [10], [11], [12], [13].

Recent in vitro experiments, on the contrary, were limited by poor availability of human myocardium. Therefore, recent work focused on analyzing the cardiodepressant effect of serum obtained from septic humans on murine [14], [15], [16], [17], [18] or rabbit [19], [20] myocardium.

In cardiac surgery, we frequently see patients with IE in need of valve surgery for acute and severe valve incompetence caused by infective processes at the level of the endocardium. From these patients, in whom intrinsic inflammation/SCMP is present, we obtained tissue specimens for our isolated right atrial myocardium model.

This study introduces an experimental set-up using human myocardium. As a particular advantage, this particular model eliminates the entirety of hemodynamic and other uncontrolled circumstances and allows establishing standardized measurement conditions. As atrial myocardium is supplied by diffusion from the (systemic) intra-atrial blood, the design of our model excludes all causes of myocardial dysfunction other than the intrinsic SCMP. This applies, in particular, to factors such as ischemic myocardial dysfunction secondary to coronary artery disease.

Moreover, there is some evidence that infective myocardial dysfunction in acute IE is not explained by (1) the development of endothelial injury or (2) the early development of atherosclerosis or (3) triggered by acute events caused by vulnerable plaques or the activation of the coagulation cascade [21].

In short, our model allows the analysis of the electromechanical properties of isolated right atrial myocardium under consistent oxygenation and nutrition conditions.

Cardiac electromechanical coupling (CEMC) measurement is a well-known technique that may be used to demonstrate changes in myocardial contractility [22]. Individual trabeculae all show a positive correlation between sarcomere length and generated tension, also referred to as the length-dependent active force, which underlies the Frank-Starling law [23]. The relation between the muscle length or sarcomere length, on the one hand, and the tension developed at optimal sarcomere length (with best contraction), on the other hand, is more pronounced in cardiac than in skeletal muscle [24].

In CEMC, the Frank-Starling mechanism (FSM) is defined as the myocardial contractility that depends on the preload, that is, the passive tension of the myocardium before active contraction. The greater the passive strain of the myocardial sarcomere is, the greater is the ensuing active contraction. Increasing the end-diastolic preload therefore leads to optimal muscle length for maximal active contraction force resulting from the optimal overlapping of actin and myosin filaments. The cardiac-specific sarcomeric structure is mainly responsible for the FSM [25]. The maximum force development of the muscle is observed at optimal length. To allow comparability, all our measurements were performed at optimal length.

Materials and methods

Human atrial myocardium was obtained from 32 patients operated upon for acute IE (group E, with microbiologically proven positive blood cultures), who required urgent or emergent aortic or mitral valve surgery caused by bacterial valve destruction with acute valve insufficiency.

Our controls (group C) consisted of atrial tissue samples from 65 patients who required valve surgery for aortic and/or mitral valve incompetence caused by noninfective causes.

The gender distribution (female/male) was similar in both groups (1/3).

Age was also not different in both groups: group E with mean age of 63.4±13.1 years (range 23–78 years) versus group C with mean age of 64.6±8.1 years (range 42–77 years; n.s.).

The human myocardium we used for our experiments consisted of the tips of the right atrial appendages that were removed in the course of venous cannulation for extracorporeal circulation. The study was in accordance with the institution’s ethical guidelines. These myocardial tissue specimens were immediately stored in warm (37°C) and oxygenated Krebs-Henseleit solution and transferred to our laboratory for CEMC measurement without delay. For this purpose, the right atrial tissue was prepared under the microscope to yield equally sized (0.3–0.5×5.0 mm) trabeculae meeting the standardized conditions for CEMC measuring using our myocardial force measurement device.

After transfer into ice-cold Krebs-Henseleit solution, the trabeculae were equilibrated in warm (37°C) oxygenated Krebs-Henseleit solution for 30 min. Being thus prepared, the right atrial trabeculae from both groups were clamped into the connecting force transducer and vibrator of the force measurement device where they were continuously flushed with warm (37°C) Krebs-Henseleit solution. Electrical stimulation was applied in square wave at a frequency of 60 beats/min. During electrical stimulation, the trabeculae clamped into the apparatus were prestreched, during continuous force measurement, from slack to optimal length, the latter being defined as the length at which maximum force output was achieved. At optimal length, the diameter of the fiber was measured under the microscope. The respective resting force and isometric force were recorded.

The passive resting force (mN/mm²) represents the passive force that was recorded after achieving optimal length of muscle stripe. This force was measured, as the response of the muscle stripe to the prestretching process, without application of electrical stimulation. The isometric force amplitude (mN/mm²) is the force amplitude recorded during supramaximal electrical stimulation of the muscle stripe at optimal length.

The active force (mN/mm²) is a calculated value that represents the difference between the passive resting force and the force amplitude (i.e. force amplitude minus resting force). To gain an insight into the relation between the degree of myocardial prestretching required, as expressed by the recorded passive resting force value, and the achievable active force response (force amplitude minus resting force) of the muscle stripe, the resting force/active force ratio was additionally calculated for each myocardial specimen.

Results

As illustrated in Table 1, the passive resting force measured at optimal length was slightly higher in the atrial myocardium of group E than group C. However, this difference did not reach statistical significance.

In Figure 1, the active force is shown as a function of the passive resting force. A linear regression line was fitted to the data for better illustration of the linear function of the relationship between the passive resting force required and the active force response in the myocardium. The slope of the linear function differed significantly between groups E and C (p<0.0001). The fact that the linear function obtained for group E is less steep than that for group C indicates that it is not possible to achieve the same level of active force response in the myocardium of patients with IE than in the myocardium of the control group.

Figure 1:

Active force (mN/mm²) as a function of passive resting force (mN/mm²) for the endocarditis group (n=32; blue) and control group (n=65; red).

All individual values form the linear regression line with significantly different slopes.

Figure 2 provides an overview of the basic findings in terms of the mean passive resting force and force values measured as well as the active force values and resting force/active force ratios calculated for groups E and C. Statistically significant differences between the mean force amplitude values measured and the active force values and resting force/active force ratios calculated for the two groups are represented in Table 1.

Figure 2:

Mean values of EMC measured [passive resting force (mN/mm²) and force amplitude (mN/mm²)] and calculated values [active force (mN/mm²) and ratio of resting force/active force] for the endocarditis group (n=32; red) and control group (n=65; blue).

Statistical analysis (Table 1) did not yield a significantly elevated passive resting force in group E, although there seems to be a trend toward a higher passive resting force in patients with IE and SCMP. Interestingly, however, the force amplitude is significantly reduced in endocarditis myocardium (4.8 vs. 5.8 mN/mm²) in SCMP. Additionally, the calculated active force was significantly lower (3.2 vs. 4.3 mN/mm²) and the passive resting force/active force ratio was significantly higher (0.53 vs. 0.37 mN/mm²) in group E.

Discussion

In view of the fact that IE is a highly complex disease in the course of which a variety of factors, including valve insufficiency and sepsis, exert their respective influences on hemodynamics, our study design compares the force-generating capacity of human atrial myocardium out of patients showing acute valve insufficiency (group C) with atrial myocardium out of patients showing acute valve insufficiency caused by bacterial vegetations (group E) in the intention to observe only the “septic” effect on myocardium.

The isolated atrial myocardium model introduces an experimental set-up that not only uses human myocardial tissue but also enables us to perform highly standardized measurements of myocardial function in the setting of SCMP. Under in vivo conditions, preload, afterload, and frequency are determined by the inflammatory process of endocarditis, so that it is not possible to make a differentiation between the hemodynamic septic effect (caused by vasodilatation) and indeed a reduced cardiac contractility (SCMP). Thus, the in vitro model of isolated human myocardium is adequate to compare the hemodynamics related to valve insufficiency (group C) with an additionally endocarditis (group E) under exactly defined conditions of preload, afterload, and frequency. Under clinical conditions, a comparison of both groups is not possible because there is no way of standardized differentiation.

Based on tissue specimens from a large patient population, our current analysis yielded three important observations: in the atrial myocardium of patients operated upon for IE, (1) the extent of force generation at optimal length was reduced, (2) the passive resting force required to achieve a given force amplitude was elevated, and (3) the passive resting force at optimal length was not significantly altered.

The passive resting force at optimal length is determined by several factors, with the parallel and serial elastic units of the muscle contributing in their capacity as muscular elements. Moreover, with any alteration of the quality and mass of connective tissue, fibrosis (perimysial or endomysial fibrosis) may be able to alter the passive resting force. On a molecular level, the number of adherent cross-bridges between the contractile units plays an important role similar to that of permanently fixed cross-bridges (latch bridges). However, as the passive resting force is similar in both groups, there is no direct evidence for alterations on the connective tissue level (inflammatory fibrosis). On the contrary, our current experiments do not exclude the possibility that the mass of connective tissue per square millimeter may be reduced. After all, an increased number of permanently fixed cross-bridges could compensate for a decreased amount of connective tissue. However, in the current literature, there is no evidence for a latch-bridge mechanism in the human myocardium of patients with IE.

The isometric force amplitude was recorded based on a standardized protocol [26], applying supramaximal stimulation, which means that we provoked the maximum achievable force response after prestretching to optimal length. In group E, the maximal force amplitude was significantly lower than in group C. This suggests (a) a reduction of contractile units contributing to force generation per square millimeter or (b) a reduced force output per contractile unit. A reduced number of contractile units per square millimeter may be the result of myocardial edema or destruction of contractile units. The destruction of contractile units in this respect may be functional (i.e. after a history of overstretching) or morphologic [i.e. as a function of insufficient supply (infarction)]. Clinically, these findings translate into an impairment of the systolic function.

The calculated resting force/active force ratio was significantly higher in group E than in group C, which indicates that in SCMP an elevated preload (resting force) is required to achieve a given active force amplitude. On the muscle preparation level, this means that there seems to be a diastolic dysfunction.

In summary, our data suggest that IE causes a severe impairment of systolic and diastolic function. Thus, we conclude that the FSM is impaired in SCMP.

The need for a higher preload (with a higher resting force), corresponding to a higher prestretching tension to achieve optimal length for maximal force development, in group E is in keeping with the FSM, stipulating the need for the optimal overlap of actin and myosin filaments in the sarcomeric structure [25] to generate maximum force. With a trend toward an elevated passive resting force in SCMP, there seems to be an attenuated diastolic dysfunction perhaps caused by fibrotic remodeling and potentially leading to an impaired systolic function caused by reduced prestretching of the muscle as a further consequence.

Therefore, we assume that SCMP has no harmful effect on the sarcomeric structure that is responsible for the maximum force development by actin and myosin interaction after optimal prestretching [25].

The diastolic dysfunction in SCMP seems to be caused by myocardial stiffness due to fibrotic remodeling of the cytoskeleton and the extracellular matrix of the myocardium. This is in keeping with recent studies reporting fibrotic tissue modifications caused by inflammation and cytokines [27]. This correlation between cytokines such as TNF-α and interleukin-10, on the one hand, and an impaired cardiac function and elevated mortality, on the other hand, has been already proven.

It is well possible that the quantity of sarcomeres in a muscle stripe of standardized muscle thickness could be reduced due to edema/swelling of myocytes or swelling of the extracellular structure. Edema could therefore be a reason for a significantly lower active force in SCMP. Of note, histopathological postmortem observations of patients who died of septic shock showed interstitial fibrosis in 100% and interstitial edema in 90% of the patients [28].

Sepsis is a circulatory compromise with microcirculatory alterations and mitochondrial damage, and all these factors reduce cellular energy production. This is the origin of sepsis-induced myocardial dysfunction mediated by cytokines [11]. This state seems to be reversible [17] in survivors, so that the myocardium may in fact experience a hibernation-like nonfunctional condition during severe sepsis. To reduce myocardial oxygen consumption, β-blocker Ivabradin and insulin are recommended as therapy strategy.

We assume that any infectious agent initially causes inflammation with circulating cytokines induced by endotoxemia. Sepsis and its hemodynamic changes including a reduced afterload caused by vascular nitric oxide (NO) result in the release of cytokines such as TNF-α and interleukins [29]. As it is also well known, TNF-α in its turn stimulates macrophages to produce NO [30]. Elevated NO and increased reactive oxygen species (ROS) lead first to myocardial hibernation and reduced contractility [29] and second to fibrotic remodeling [31], [32], [33] by tipping the balance between ROS and antioxidant enzymes.

The early use of antibiotics to keep inflammatory mediators from exerting their full effects may be considered recommendable. Critical care monitoring of troponin and B-type natriuretic peptide (BNP) is crucial for the detection of SCMP, which in the course of time will result in cardiac failure and left ventricular dilatation [23].

Nevertheless, an elimination of the septic focus (bacterial vegetation) in good time by early operation should be recommended [34].

Limitations

The study protocol was very simple and focused on mechanical performance. Additional histopathological data of the preparations are not available. Although the current data suggest an impairment of myocardial function in atrial myocardium, it is likely that similar effects could be observed in ventricular myocardium, although this cannot be derived with a sufficient degree of certainty from our experiments, which were implemented using atrial myocardium only. In this respect, the differences between atrial and ventricular myocardium in terms of their blood supply, microcirculation, and pressure conditions as well as other characteristics must be considered.

Furthermore, the proven cardiac depressant effect is not independently observed in septic patients, as patients with IE undergoing cardiac surgery (as an essential requirement to obtain human myocardium for our experimental set-up) mostly also present with valve insufficiency.

Perspectives

The data argue for the presence of a significant cardiodepressant effect on atrial myocardium associated with the occurrence of endocarditis. There are several candidates that may contribute to the pathophysiological and pathoanatomical chain leading to reduced contractility. However, based on the current data, no decision in favor of one or some of these factors is justified.