Intensive Care Medicine

, Volume 29, Issue 10, pp 1802–1807

The calcium sensitizer levosimendan attenuates endotoxin-evoked myocardial dysfunction in isolated guinea pig hearts


    • Klinik für Anästhesiologie und IntensivmedizinUniversitätsklinikum Essen
  • Jürgen Peters
    • Klinik für Anästhesiologie und IntensivmedizinUniversitätsklinikum Essen

DOI: 10.1007/s00134-003-1879-8

Cite this article as:
Behrends, M. & Peters, J. Intensive Care Med (2003) 29: 1802. doi:10.1007/s00134-003-1879-8



Sepsis-evoked myocardial dysfunction is possibly due to decreased myofilament calcium sensitivity, and a calcium sensitizer may thus specifically improve contractility in sepsis by enhancing myofilament calcium sensitivity. We examined whether the calcium sensitizer levosimendan mitigates myocardial dysfunction and improves contractility in hearts isolated from endotoxin-treated guinea pigs.

Design and setting

Prospective, controlled, randomized animal study in a university research laboratory.


Guinea pig hearts isolated 4 h (n=10) or 18 h (n=8) following E. coli LPS (4 mg/kg i.p.) and hearts from sham-treated controls (n=11 and n=6).


Isolated hearts were perfused at constant aortic pressure [Krebs-Henseleit buffer, heart rate: 300/min, left ventricular (LV) diastolic pressure: 6–8 mmHg], and LV developed pressure (LVdP) and LVdP/dt were continuously assessed. Levosimendan was added to the perfusate in incremental concentrations (0.03, 0.1, 0.3 µM).

Measurements and results

Endotoxin resulted in a significant decrease in LVdP by 20±6% and 43±8%, in +LVdP/dt by 16±5% and 44±7%, and in −LVdP/dt by 27±8% and 47±8% after 4 and 18 h, respectively. In septic hearts levosimendan increased LV function concentration-dependently by 32±4% (LVdP), 33±5% (+LVdP/dt), and 37±7% (−LVdP/dt) 4 h and by 31±6% (LVdP), 33±6% (+LVdP/dt), and 32±7% (−LVdP/dt) 18 h after LPS. However, levosimendan increased myocardial function similarly in control hearts.


While the calcium sensitizer levosimendan markedly improved LV contractility in hearts from both endotoxic and sham animals, it failed to specifically abolish endotoxin-evoked myocardial dysfunction. Thus, decreased calcium sensitivity either does not play a major role in endotoxin-evoked cardiomyopathy or the location of its pathomechanism differs from levosimendan's site of action.


SepsisCardiomyopathyMyocardial functionContractilityCalcium sensitivityLangendorff


Sepsis and the systemic inflammatory response syndrome evoke myocardial dysfunction. Although profound systemic vasodilatation often conceals cardiac dysfunction, left ventricular (LV) contractility is diminished, diastolic relaxation is impaired, and ventricular dilatation occurs [1]. This "septic cardiomyopathy" contributes to the high mortality in sepsis [2]. While the immunological cascade triggered by sepsis and endotoxemia evoking myocardial dysfunction has largely been uncovered [3], controversy continues about the responsible subcellular mechanisms. Decreased myoplasmic calcium transients in septic hearts may play a role [4, 5], but other investigations suggest a decrease in myofilament calcium sensitivity [6, 7, 8, 9, 10, 11, 12].

Septic cardiomyopathy is treated with catecholamines [13], but the response to β-adrenergic stimulation is diminished [14], and the invoked increase in intracellular cAMP concentrations may amplify the suspected intracellular calcium overload in sepsis [15], promoting life-threatening arrhythmias. Calcium sensitizers, in contrast, due to their specific mechanisms of action, may ameliorate myocardial dysfunction in sepsis while avoiding the disadvantages of catecholamine therapy. Furthermore, if myocardial dysfunction were the consequence of decreased myofilament calcium sensitivity, the positive inotropic effects of calcium sensitizers should be enhanced in septic hearts. However, this issue has received little attention. The calcium sensitizer MCI-154 improved indices of myocardial contractility when given 10 h after endotoxin injection in rabbits [16], but it has distinct phosphodiesterase III inhibiting properties that could explain this effect. Furthermore, while the novel calcium sensitizer levosimendan improved cardiac output in endotoxemic pigs [17], its effect on myocardial contractility is unknown.

Therefore to test the hypothesis that levosimendan improves endotoxin-evoked myocardial dysfunction irrespective of systemic hemodynamic alterations we assessed its effects on myocardial contractility in hearts isolated from endotoxin-treated guinea pigs.


Instrumentation and procedures

Male Duncan Hartley guinea pigs weighing 350–450 g (Harlan-Winkelmann, Borken, Germany) were supplied with commercial guinea pig chow and tap water ad libitum before and after endotoxin administration. The experimental protocol was approved by the local animal care committee and was in accordance with the guidelines of the American Physiological Society. Following either lipopolysaccharide (LPS) or sham pretreatment guinea pig hearts were isolated during isoflurane anesthesia 10 min after injection of sodium heparin (1,000 IU intraperitoneally). Hearts were cooled in situ with ice-cold saline (4°C), and a cannula was placed into the ascending aorta. Aortic pressure was maintained at 60 mmHg, and perfusion of the coronary arteries commenced in a nonrecirculating Langendorff technique using a freshly prepared, modified Krebs-Henseleit bicarbonate buffered solution (NaCl 120 mM, KCl 4.0 mM, NaHCO3 24 mM, CaCl2 2.0 mM, MgSO4 1.2 mM, KH2PO4 1.0 mM, glucose 10 mM), bubbled with 95% O2, 5% CO2, and yielding pH 7.4, PO2 550 mmHg, and PCO2 38 mmHg. Hearts were suspended in a temperature-controlled chamber and myocardial temperature was maintained at 36.9±0.1°C. Right atrial pacing (300/min) was continued throughout the experiments.

Levosimendan was added to the perfusate with a syringe pump connected to a sideport of the Langendorff apparatus. Levosimendan was synthesized by Orion-Farmos Pharmaceuticals; all other chemicals were analytical grade and purchased from Sigma Chemical.


Aortic perfusion pressure (electromanometry, PvB DPT-6003, Kirchseeon, Germany) and coronary perfusate flow (ultrasound transit time flow probe, Transonics, Ithaca, N.Y., USA) were continuously measured in the perfusion line, close to the aortic cannula. Myocardial temperature was measured using a thermocouple probe (Physitemp IT-23, Clifton, N.J., USA) placed in the right ventricle. Contractile function was assessed by measuring LV developed pressure (LVdP), using a saline-filled latex balloon (Hugo Sachs, March/Hugstetten, Germany) inserted into the left ventricle via the left atrium. LV diastolic pressure was maintained at 6–8 mmHg. +LVdP/dt and −LVdP/dt, i.e., the maximum rate of LV pressure development and relaxation, respectively, were derived electronically from the digitized LVdP signal to assess systolic and diastolic performance. Variables were continuously recorded on a digital thermoarray chart recorder (Astro-Med Dash 8U, West Warwick, R.I., USA).

Experimental protocol

Following randomization guinea pigs received either Escherichia coli lipopolysaccharide (LPS; 4 mg/kg intraperitoneally; serotype O127:B8, Sigma-Aldrich, Taufkirchen, Germany) or an equivalent volume of normal saline. In the first series hearts were isolated 4 h after LPS (n=11) or saline (n=10). In subsequent experiments hearts were isolated 18 h after LPS (n=6) or saline (n=8). Following stabilization of variables aortic perfusion pressure and coronary flow, LV pressure, and myocardial temperature were measured at baseline and continuously thereafter. Levosimendan was applied in incremental concentrations of 0.03, 0.10, and 0.30 µM, reflecting concentrations in a therapeutic clinical setting [18]. Variables were assessed following an equilibration phase of at least 8 min.

Data analysis

Data are reported as means ±SEM. A two-way analysis of variance followed by post hoc tests (one-way repeated-measurements analysis of variance and unpaired t tests) accounting for the time course of experiments and the experimental groups were applied to compare values of variables, using Bonferroni's adjustment of the α error. An α less than 0.05/n was considered statistically significant for rejecting the null hypothesis, i.e., that there is no difference among experimental groups and different levosimendan concentrations.


Guinea pigs receiving LPS developed signs of sepsis such as diarrhea, piloarrection, and lethargy whereas sham animals did not. No animal died as a consequence of LPS or saline administration during the 4- or 18-h observation periods. Endotoxin pretreatment significantly decreased by 20±6% and 43±8% LV developed pressure after 4 h (Fig. 1) and 18 h (Fig. 2), respectively, vs. the respective sham group. Furthermore, LVdP/dt also markedly decreased after 4 h (+LVdP/dt −16±5%, −LVdP/dt −27±8%) and 18 h (+LVdP/dt −44±7%, −LVdP/dt −47±8%). Coronary flow at constant aortic pressure decreased significantly only after 18 h (−16±9%).
Fig. 1.

Left ventricular (LV) developed pressure, coronary flow, and maximum rate of LV pressure development and relaxation with increasing levosimendan concentrations in perfused guinea pig hearts isolated 4 h after intraperitoneal LPS (LPS 4 h, n=10) or saline (sham 4 h, n=11) injection. Means ±SEM. Two-way analysis of variance followed by post hoc tests was used to evaluate differences between groups (LPS and sham) and levosimendan concentrations. *p<0.05 vs. baseline of the respective group

Fig. 2.

Left ventricular (LV) developed pressure, coronary flow, and maximum rate of LV pressure development and relaxation with increasing levosimendan concentrations in perfused guinea pig hearts isolated 18 h after intraperitoneal LPS (LPS 18 h, n=8) or saline (sham 18 h, n=6) injection. Means ±SEM. Two-way analysis of variance followed by post hoc tests was used to evaluate differences between groups (LPS and sham) and levosimendan concentrations. *p<0.05 vs. baseline of the respective group

Levosimendan increased LVdP, LVdP/dt, and coronary flow in a concentration-dependent fashion in the endotoxin-treated hearts (Figs. 1, and 2). At 4 h after endotoxin pretreatment levosimendan increased LVdP by 32±4%, +LVdP/dt by 33±5%, −LVdP/dt by 37±7%, and coronary flow by 29±6%. At 18 h after endotoxin pretreatment, levosimendan also markedly increased LVdP by 31±6%, +LVdP/dt by 33±6%, −LVdP/dt by 32±7%, and coronary flow by 13±2%. The increases in LVdP and LVdP/dt seemed to approach a plateau at a levosimendan concentration of 0.1 µM, whereas coronary flow continued to increase with higher levosimendan concentrations.

A similar improvement in LV function with levosimendan was seen in sham hearts (Figs. 1, 2). At 4 h after the sham procedure levosimendan increased LVdP by 22±5%, +LVdP/dt by 32±7%, −LVdP/dt by 32±8%, and coronary flow by 28±8%. After 18 h levosimendan increased LVdP by 15±6%, +LVdP/dt by 25±8%, −LVdP/dt by 18±7%, and coronary flow by 13±2%. Again, levosimendan's concentration-dependent effect on LVdP and LVdP/dt was most pronounced at concentrations up to 0.1 µM, whereas coronary flow continued to increase with higher levosimendan concentrations.

However, the positive effects of levosimendan on LV pressure development and relaxation as well as on coronary flow did not significantly differ between sham and endotoxin-pretreated animals. Furthermore, there was no difference between the 4-h and the 18-h groups.


The calcium sensitizer levosimendan improves substantially myocardial contractility and lusitropy in a concentration-dependent fashion both in endotoxin-exposed and in normal hearts, but calcium sensitization with levosimendan does not appear to specifically abolish endotoxin-evoked LV dysfunction.

Intraperitoneal endotoxin injection results in vivo in nervous system depression, minor changes in arterial blood gas tensions, and a marked decrease in arterial blood pressure [19]. Mortality after 18 h is reported to be 20–30% but varies considerably among investigations. The endotoxin dose used is known to evoke myocardial dysfunction in guinea pigs within 1–4 h, that is most pronounced after 12–16 h, and persists for several hours after isolation of the heart [19, 20]. Hearts were isolated both 4 and 18 h after endotoxin to investigate whether the length of incubation affects myocardial dysfunction or levosimendan's potential effects, as the pathomechanisms of early and late myocardial dysfunction may differ [3].

Indeed, pretreatment with LPS for 18 h resulted in a more distinct myocardial dysfunction, although comparability of 4 and 18 h baseline data is somewhat limited. In the 18-h groups baseline myocardial function in the sham group was better than in the sham group at 4 h which included some hearts with poor baseline function not taken out of analysis. Levosimendan's positive inotropic effects, however, were not altered by the incubation period or by baseline function. In fact, although starting from different baselines, the absolute increases in LV developed pressure and LVdP/dt were almost identical in all LPS and sham groups, suggesting that the specific mechanisms for endotoxin-evoked myocardial dysfunction were not specifically interfered with by levosimendan.

Whether endotoxin-evoked myocardial dysfunction results from decreased calcium transients or decreased myofilament responsiveness to calcium remains controversial. We hypothesized that if an endotoxin-induced decrease in calcium sensitivity were the responsible mechanism, contractility should increase to a greater extent after levosimendan in endotoxin-pretreated hearts than in sham hearts. However, the lack of such an effect does not fully exclude decreased myofilament responsiveness to calcium following endotoxin as the subcellular defect leading to decreased calcium sensitivity could be independent from Ca2+ binding to troponin C and therefore differ from levosimendan's site of action. Other investigations indicate that myocardial dysfunction in endotoxin-treated guinea pigs is caused by decreased calcium transients [4] and an intracellular calcium overload [21]. These results are consistent with similar effects of levosimendan in both endotoxic and normal hearts observed in our study since predictably these hearts should benefit to a similar extent from increased calcium sensitivity.

Levosimendan is the first calcium sensitizer approved for therapy of severe heart failure [22]. Since decreased myofilament calcium sensitivity has repeatedly been demonstrated in experimental septic cardiomyopathy [6, 7, 8, 9, 10, 11, 12], it has been speculated that calcium sensitizing agents may also ameliorate myocardial dysfunction of sepsis [12]. However, effects of calcium sensitizers in septic cardiomyopathy have not been thoroughly investigated. The calcium sensitizer MCI-154 improves cardiac function during endotoxic shock in rabbits when given 10 h after intravenous administration of E. coli endotoxin [16]. While its calcium-sensitizing effect was confirmed in myocardial fibers isolated from rats with endotoxic shock [23] MCI-154 has, in addition to its calcium-sensitizing effect, even more pronounced phosphodiesterase III inhibiting properties [24, 25] that could explain its reported inotropic effects. In fact, phosphodiesterase III inhibitors are known to improve myocardial function in both patients and animals with sepsis [26, 27, 28]. Accordingly, the effect of calcium sensitization on contractility can hardly be addressed by using MCI-154. Levosimendan, in contrast, acts preferably as a calcium sensitizer, although phosphodiesterase inhibiting properties may contribute to its positive inotropic effect, especially at higher concentrations [25, 29, 30, 31]. By demonstrating levosimendan's inotropic effects in isolated hearts our experiments also explain in vivo studies in pigs where an increased cardiac index at decreased systemic vascular resistance and blood pressure has been reported with levosimendan [17].

Increased calcium sensitivity was not associated with impaired relaxation, due to levosimendan's calcium-dependent binding to troponin C. In our experiments levosimendan even increased myocardial lusitropy concentration-dependently, possibly due to its phosphodiesterase inhibiting properties. Consequently levosimendan improves systolic contractile function by calcium sensitization while potential deleterious effects of calcium sensitization on diastolic function are avoided. Thus, due to its demonstrated positive inotropic effects in the context of endotoxemia levosimendan may be a rational addition to catecholamine therapy in sepsis-evoked myocardial dysfunction.

Coronary blood flow decreased in LPS hearts, indicating that endotoxin-pretreatment did not completely abolish coronary autoregulation, although distinct coronary vasodilatation is frequently seen in sepsis [32]. With levosimendan coronary flow increased dose-dependently in all groups. However, the effect was most pronounced at higher levosimendan concentrations, suggesting that coronary dilatation is not the consequence of increased myocardial metabolism due to increased myocardial work but of opening of KATP channels with higher levosimendan concentrations [31].

In summary, the calcium sensitizer levosimendan markedly improves myocardial contractility and lusitropy with septic cardiomyopathy. However, since levosimendan improves myocardial function in endotoxic and normal hearts to a similar extent, it does not specifically abolish endotoxin-evoked myocardial dysfunction. As levosimendan's mode of action, i.e., increased myofilament calcium sensitivity, addresses a potential mechanism of endotoxin-evoked myocardial dysfunction, our results also suggest that decreased calcium sensitivity either does not play a major role in endotoxin-evoked cardiomyopathy, or that its pathomechanism differs from levosimendan's site of action.

Copyright information

© Springer-Verlag 2003