Cardiovascular response to dobutamine stress predicts outcome in severe sepsis and septic shock

The cardiovascular profile of septic shock has been well defined over the last 30 years. One of the major areas of investigation has been the cardiovascular correlates of outcome. Early clinical studies, prior to the widespread adoption of the balloon-tipped pulmonary artery flotation catheter with thermodilution capability in the 1980s, tended to suggest that fatal sepsis and septic shock were characterised by a low cardiac output (CO) and high systemic vascular resistance (SVR) [1]. However, the concept of septic shock as a hyperdynamic circulatory state characterised by high cardiac output and low systemic vascular resistance eventually became well established [2, 3]. Several baseline cardiovascular variables have been proposed to differentiate between survivors and non-survivors of septic shock including heart rate (HR), CO, SVR and oxygen delivery index (DO₂I) [4‐7]. Studies using radionuclide cineangiography and volumetric echocardiography have further demonstrated an association between survival and biventricular dilatation and ejection fraction (EF) depression [3, 8‐10]. Evidence has also accumulated suggesting that fatal septic shock is associated with decreased responsiveness to fluid resuscitation [11, 12].

Bedside experience and several clinical studies suggest that blunted cardiovascular responsiveness to catecholamine support is also a marker of high risk of death in sepsis and septic shock [13]. However, the relationship of ventricular ejection fraction and/or ventricular volume responses and other ventricular volume-related cardiovascular variables under dobutamine stress to survival in sepsis and septic shock is undefined. This study was designed to assess ventricular and cardiovascular performance, oxygen transport, and oxygen delivery responses during dobutamine challenge and their association with survival in sepsis.

This study received Institutional Review Board approval. Informed consent was obtained from all subjects or appropriate representatives prior to enrolment.

Adult (more than 18 years of age) Intensive Care Unit patients were recruited into this prospective, non-randomised, open interventional study. In order to be eligible, patients were required to have developed severe sepsis or septic shock within the preceding 48 hours and to have both a pulmonary artery catheter and arterial pressure catheter in place (based on clinical requirements as defined by the attending physician). Study eligibility was further contingent on confirmation of correct placement of the pulmonary artery catheter based on appropriate pressure traces on insertion and position on chest radiography.

The presence of severe sepsis was defined by 1991 SCCM/ACCP Consensus Statement on Sepsis Definitions [14]. These criteria included: the presence of suspected or documented infection; three of the following four systemic inflammatory response elements temperature >38°C or <36°C, heart rate >90 beats/min, respiratory rate >20 breaths/min or PaCO₂ <32 mm Hg and white blood cell count >12,000 cells/mm³, <4,000 cells/mm³ or >10 per cent immature (band) forms; and evidence of at least one organ failure.

Septic shock was defined by the presence of post-fluid resuscitation hypotension (mean arterial pressure <65 mmHg) necessitating vasopressor support.

Patients with known ischemic cardiac disease and mixed shock states with clinically significant primary cardiogenic, haemorrhagic or pulmonary embolic elements were excluded.

After consent was obtained, each patient underwent graded dobutamine challenge as follows: If required, subjects initially received saline infusion to a pulmonary artery wedge pressure (PWP) of 15 mmHg as required because pilot studies had shown the potential for significant hypotension with dobutamine absent adequate intravascular volume as evidenced by PWP of 15 or higher. After documentation of a PWP of 15 mmHg or higher, baseline haemodynamic and vital parameters (on clinically indicated doses of vasopressors) were obtained. These included heart rate (HR), systemic arterial pressures, central venous pressure (CVP), pulmonary artery pressure (PAP), pulmonary wedge pressure (PWP) and CO. Thermodilution cardiac outputs were measured by the averaging of three of four successive injections of 10 mL of cold (6 to 10°C) dextrose 5 per cent in water at end-expiration (with the outlying values discarded). Recorded values for PAP, PWP and right atrial pressure (RAP) were also obtained at end-expiration from graphic recordings. Right and left ventricular ejection fractions (EFs) were assessed through first-pass gated technetium⁹⁹ cineangiography as described below. In addition, simultaneous arterial and mixed venous blood samples (distal pulmonary artery port) were obtained for analysis of the serum pH, partial pressure of oxygen (pO₂), partial pressure of carbon dioxide (pCO₂), and oxygen saturation (Bayer-CIBA Model 865, Tarrytown, NY).

After baseline values were obtained, a graded dobutamine challenge was performed. Dobutamine infusion through the pulmonary artery catheter introducer port was initiated at 5 μg/kg/min. After 15 minutes, repeat non-invasive and invasive haemodynamic parameters were obtained over approximately five minutes. At this time, subjects also underwent repeat first pass radionuclide assessment of biventricular ejection fractions and arterial and mixed venous blood gas analysis. The dobutamine infusion rate was increased by 5 μg/kg/min at 20 minute intervals to a maximum of 15 μg/kg/min with the assessment process repeated during the final five minutes of each interval. During the infusion, fluid and additional vasopressor support (beyond baseline needs) to maintain a minimum PWP of 15 mmHg and a mean arterial pressure (MAP) of 65 mmHg was mandated; however, no subject required such support. During the period of the study environmental stimulation was minimised. All subjects were continuously monitored with continuous electrocardiographic (ECG) monitoring and pulse oximetry according to standard protocol in the intensive care unit. The protocol required discontinuation of dobutamine if any transient adverse effects were noted including significantly increased ventricular ectopy or decreased blood pressure. After study completion, dobutamine was tapered over 20 minutes and then discontinued. The patient's subsequent intensive care and hospital course were directed by normal clinical practice.

Dobutamine, a racemic synthetic catecholamine composed of two enantiomers with divergent sympathetic activity, was developed as an inotropic agent and is most commonly used for support of patients with congestive heart failure and other low output states of cardiogenic decompensation [17]. A series of studies have assessed this compound as a potential adjunctive therapy in severe sepsis for purposes of augmentation of oxygen delivery [18‐21]. In addition, a dobutamine stress test with determination of oxygen consumption responses to assess survival potential of patients with severe sepsis and septic shock has been examined [22‐24]. No studies have assessed volumetric responses to graded dobutamine infusion in septic patients in relationship to outcome.

This study, like that of Jellema and colleagues [25], shows that dobutamine infusion in patients with septic shock generates a highly heterogenous cardiovascular response. We also demonstrate that dobutamine-stressed ventricular and cardiovascular responses are highly correlated with outcome. The observation that greater increases in left ventricular contractility (LVEF, sBP/LVESVI, LVSWI) and right ventricular lusitropy (RVEDVI, RVEDVI/RAP) indices are associated with survival in severe sepsis and septic shock in univariate analysis, is entirely novel. In addition, an independent contribution to prediction of survival under dobutamine stress from cardiovascular performance (SVI) and left ventricular contractility (sBP/LVESVI) responses has also not been previously reported. Notably, the correlative strength of these responses is substantially more powerful than that generated by initial HR, the only baseline hemodynamic index to predict outcome.

These results are important for at least three reasons. First, they support the notion that simple haemodynamic responses to dobutamine stress can potentially differentiate between survivors and non-survivors of sepsis and septic shock. This confirms long-standing clinical observations and suggests that these responses may be useful in terms of clinical prognostication for purposes of risk stratification in clinical trials. Second, they suggest that dynamic cardiovascular responses may be superior to static performance parameters with respect to their association with survival. Finally, these observations provide important insights regarding the mechanism underlying the cardiovascular response to adrenergic stimulation in survivors versus non-survivors of septic shock.

The observation that only HR among all baselines haemodynamic/oxygen metabolism variables assessed was significantly correlated with outcome has precedence in other studies of septic shock [4, 26]. Azimi and coworkers demonstrated that only heart rate among initial haemodynamic variables was associated with outcome with a value of less than 106 beats/min linked to survival [26]. No other initial parameters were associated with prognosis. Parker and colleagues similarly demonstrated that HR of less than 95 beats/min at presentation was associated with survival in a study of 48 patients with septic shock [4]. There are several possibilities to explain the association of an increased HR with a higher probability of death. The increased HR in non-survivors may indicate a higher level of sepsis-induced cardiovascular stress due to a more severe disease state. Alternately, it may represent a secondary compensatory response as a consequence of decreased sensitivity to the direct myocardial inotropic effects of endogenous catecholamines in patient with particularly severe septic myocardial dysfunction [3, 27]. Given that baseline SV and sBPVESI was similar in both groups, this possibility seems less likely. Another possibility is that non-survivors have increased sensitivity to catecholamine stimulation. This also seems unlikely since several studies have shown β-adrenergic uncoupling in myocardial tissue exposed to inflammatory mediators such as those found circulating in patients with septic shock [28, 29]. Since the study was not designed to determine the cause of this finding, no definitive conclusion can be drawn.

Despite observations dating back several decades suggesting that reduced response to catecholamine support in septic shock is highly predictive of poor outcome, most recent work examining the prognostic significance of dobutamine stress in this situation has focused on oxygen consumption/delivery indices [22‐24]. Both Vallet et al [24] and Rhodes et al [22] have shown that increases of oxygen consumption of greater than 15 per cent with 10 μg/kg/min during severe sepsis are associated with a favorable outcome. However, it is not at all clear that it is necessary to invoke oxygen consumption/delivery responses for predictive power. Both studies clearly demonstrated that cardiac index responses were also highly associated with outcome. However, neither closely examined cardiac performance parameters as correlates of outcome.

In our study, cardiovascular performance as indicated by SVI response was the dominant correlate to outcome with a single cut off value of 8.5 mL/min able to differentiate 21 of 23 outcomes. The predictive power of most other parameters appeared to be physiologically linked to this response and did not persist in multivariate analysis (CI, PP, OER, PAD, SVRI, PVRI, DO₂I, MvO₂, CvO₂). sBP/LVESVI, a relatively load-independent index of left ventricular contractility [30], was the only variable that possessed predictive power independent of stroke volume index (in analysis with both absolute and relative changes from baseline). This response appears to statistically dominate other contractility indices (LVEF and LVSWI) in terms of contribution to the model. If sBP/LVESVI is removed from the model, then LVEF becomes a significant contributor along with SVI.

The fact that both cardiac performance and contractility increases in response to dobutamine stress are highly associated with outcome in septic shock is consistent with the possibility that the presence of cardiovascular reserve (that is an indication of submaximal cardiovascular response to baseline stress) may be critical to survival. This concept of preserved physiologic reserve as a marker of good prognosis in septic shock has previously been postulated by Hayes et al [31] and Rhodes et al [22] albeit in the context of oxygen consumption indices. However, it may relate directly to cardiovascular performance and contractility indices as shown in this study.

This study may also demonstrate that the concept of preload reserve used most commonly in reference to ischemic heart disease associated limitations in cardiac performance may also apply to septic shock [32]. Several groups of investigators have shown that subjects with clinically asymptomatic impairment of systolic cardiac function can be differentiated from symptomatic patients by the presence of exercise stress-induced increases in SVI in association with ventricular dilatation [33, 34]. No direct link has yet been made between survival and the presence of this phenomenon among patients with impaired heart function. Despite this, our study may be unmasking a parallel, related phenomenon in that both exercise and dobutamine challenge represent stressed states in comparison to baseline.

Similarly, our observations of an association of right ventricular dilatation (Table 2, Figure 5a) in response to dobutamine (with a similar trend for the left ventricle) appear to be related to our previous observations which demonstrate a relationship between right and left ventricular dilation and survival in septic shock [3, 8, 35].

One possibility to tie these observations together is that recruitment of preload reserve is a critical but late response during high degrees of cardiovascular stress. If that is the case, then a lack of ability to recruit ventricular volume at any point during cardiovascular stress may indicate that the subject is already operating at the maximum efficacy of the Frank-Starling response and has no additional cardiovascular reserve. In this formulation, patients with chronic systolic dysfunction would be more likely to be symptomatic, while patients with septic shock would be more likely to die if such reserve was limited. The nature of the limitation of cardiovascular reserve in conditions of high stress would be a combination of the baseline cardiovascular fitness and the degree of acute cardiovascular stress.

The study demonstrates that cardiovascular hypo-responsiveness to catecholamine stimulation is strongly associated with mortality in severe sepsis and septic shock. Compared to baseline differences in heart rate, dobutamine stress-induced variations in stroke volume responses are much more closely related to outcome. Dobutamine-induced differences in contractility responses also appear to possess independent correlative power with respect to mortality. One physiological mechanism supporting increased cardiac output appears to involve ventricular dilatation in survivors suggesting an important role for "preload reserve" in optimal cardiovascular response to severe sepsis and septic shock. This response may be related to previous observations demonstrating that increased ventricular dilatation is characteristic of survival from septic shock.

This study makes several novel observations. However there are significant limitations to the results. First, despite the fact that dobutamine-stress induced cardiovascular responses are highly associated with outcome, their utility for risk stratification as related to research or clinical prognostication remains unproven. A prospective study would be required to validate their utility for that purpose. Second, these data do not suggest that any of the variables associated with outcome drive these are responsible for these outcomes. It is much more likely that these associations are correlations without causality. For example, reduced stroke volume response to dobutamine likely indicates underlying severity of illness rather than being a direct contributor to risk of death. Finally, there may be significant unmeasured confounders for some of the documented responses. Higher mean intra-thoracic pressures due to more severe lung disease and greater ventilatory requirements could contribute to the smaller RVEDVI and RVEDVI/RAP responses to dobutamine. These patients could also be at higher risk of death. Unfortunately, mean airway pressures are not available for analysis.

Although a variety of dobutamine-stimulated cardiovascular and ventricular responses are associated with outcome in sepsis and septic shock, SVI increases are most strongly correlated with survival. Since this stimulation test involves transient administration of dobutamine, this response is probably indicative of preservation of cardiac signalling mechanisms and of cardiovascular reserve in the septic patient. Notably, the predictive strength of these responses is substantially more powerful than that generated by initial HR, the only baseline haemodynamic index associated with outcome.

Further research is required to better understand the mechanisms underlying differing cardiovascular responses during severe sepsis and septic shock both at baseline and under exogenous catecholamine stress.

1) Survival from sepsis and septic shock is closely associated with dobutamine-stimulated increases in stroke volume.

2) Survivors of sepsis and septic shock are characterised by preserved cardiovascular reserve including preload reserve.