­

Sepsis remains a significant health challenge affecting more than 750 000 patients each year in the USA. In recent times, there has been an increasing recognition and understanding of cardiac involvement in sepsis.1 2 Septic cardiomyopathy is described as acute-onset reversible cardiac dysfunction developing in the setting of sepsis, which may present as systolic and or diastolic dysfunction of one or both ventricles in the absence of an ischaemic event.2 Prior studies have largely focused on left ventricular involvement; however, in recent times, there has been greater recognition of right ventricular (RV) involvement in sepsis.1 In this issue of Heart, Kim and colleagues present a timely study looking at the impact of RV dysfunction on the clinical outcomes of septic shock.3 This single-centre retrospective study performed transthoracic echocardiography (TTE) <72 hours of admission in patients with septic shock. They excluded patients with congenital heart disease, symptomatic heart failure, moderate or greater valvular heart disease, and prior TTE abnormalities. The primary outcome of 28-day mortality was higher in patients with septic cardiomyopathy (35.9% vs 26.8%, p<0.01), particularly in patients with isolated RV dysfunction (5.2% vs 2.2%, p=0.04) and combined left ventricular systolic and RV dysfunction (8.6% vs 4.6%, p=0.03). Multivariate analysis showed RV dysfunction to be independently associated with higher 28-day morality. We commend the authors for their work addressing important questions and seek to discuss two major issues.

Defining right ventricular dysfunction

The study cohort and definition, and TTE timing and technique have led to a varying prevalence of RV dysfunction in literature, ranging from 17.8% to 79.0% (online supplementary table 1). The definitions have included two-dimensional measurements or ratios thereof, regional systolic assessment, global systolic assessment, semiquantitative visual assessment, multimodal approach, strain imaging and RV cardiac output. The complex geometry of the RV, further confounded by mechanical ventilation, position or bandages in critical illness, make consistent measurement of all these parameters extremely challenging. All these techniques have unique limitations. Regional systolic measurements are easily obtained, less dependent on image quality and reproducible, but are angle dependent and assume a single region as representative of global function.4 RV fractional area change requires clear endocardial delineation, which is obtained only in a minority of the patients. Semiquantitative assessment of RV size and function is as accurate as quantitative parameters in predicting outcomes in patients with sepsis but lacks objectivity.1 Global longitudinal strain remains an advanced technique requiring proprietary software and imaging expertise, and lacks a normal reference value, all of which limit its applicability in contemporary ICU practice.4 The presence of reduced RV ejection fraction measured using the pulmonary artery catheter was able to predict worse outcome in sepsis with higher accuracy; however, the timing and utility of using both TTE and pulmonary artery catheterisation in RV dysfunction in sepsis need further elucidation, given the lack of robust data from clinical trials for the latter.5

In addition, the frequency of echocardiography in the study of sepsis and septic cardiomyopathy remains unanswered. Vieillard-Barron et al noted an additional 21% to develop systolic dysfunction after 24–48 hours, in addition to 31% that developed on admission, alluding to the dynamic nature of septic cardiomyopathy.6 As seen in online supplementary table 1, the prevalence of RV dysfunction is higher in studies where echocardiography was performed at 48–72 hours compared with studies which enrolled within 24 hours. Lastly, compared with this particular study, previous work from our group performed over a relatively similar time period excluded a much larger population of patients due to pre-existing or risk factors for RV dysfunction.1 Left ventricular dysfunction remains the most common cause of RV dysfunction (online supplementary table 1). The presence of asymptomatic left ventricular dysfunction may have put these patients at a higher risk of developing RV dysfunction, or the TTE may have ‘unmasked’ latent cardiomyopathy in the intensive care unit. This highlights significant variability in the exclusion criteria, which need to be carefully evaluated when determining the true spectrum of this problem.

Mortality with RV dysfunction

There is increasing interest in understanding the mechanisms associated with septic cardiomyopathy (figure 1). The RV is a thin-walled structure supplying blood to a high-compliance, low-resistance pulmonary circulation. Sepsis is characterised by vasodilation and decreased resistance in systemic circulation, whereas the pulmonary vasculature resistance may increase due to hypoxic vasoconstriction, acute lung injury, hypercarbia, imbalance in endogenous vasodilators and vasoconstrictors, vascular microthrombi and mechanical ventilation.7 Fluid administration, the cornerstone of sepsis management, further increases the RV preload, which in the setting of elevated afterload compensates by further increasing end diastolic volume and pressure, leading to isolated RV dysfunction. Lastly, coronary artery inflammation and plaque rupture in sepsis may be associated with undiagnosed acute myocardial infarction in this population, impairing contractility.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Mechanism of RV dysfunction associated organ failure and mortality in sepsis. RV, right ventricular.

The severity of illness, positive fluid balance, vasoactive medications and need for mechanical ventilation tend to be higher in non-survivors compared with survivors. This raises the question if the finding of RV dilation and dysfunction in sepsis and its association with poor outcome is a marker of severity of sepsis-necessitated resuscitation. Lack of customising resuscitation strategies in RV dysfunction may lead to its failure.5 Furthermore, RV dysfunction is associated with false-positive fluid responsiveness, wherein changes in left ventricular outflow parameters or its surrogate with respiratory variation are used to determine fluid responsiveness.8 They may be falsely positive from exaggerated ventricular interdependence in RV failure rather than the ventricle functioning on the steep portion of the Frank-Starling curve, leading to inadvertent fluid loading in an already compromised RV. Additionally, increased preload and passive renal vein congestion, in addition to decreased forward flow and prerenal arterial hypoperfusion, may result in acute kidney injury and cardiorenal syndrome. Therefore, positive fluid balance, comorbidities, need for mechanical ventilation, acute kidney injury and concomitant myocardial infarction, which are known to be associated with higher in-hospital mortality, may explain the worse clinical outcomes with RV dysfunction.

In summary, there is a crucial need to understand the how to identify RV dysfunction in sepsis and the causative mechanisms associated with higher mortality in this population, which will significantly influence how we prevent and manage this disease process.