Population study

We prospectively included 60 consecutive patients (mean age 62 ± 16 years; 45 men) admitted from November 2008 to March 2010 to the Intensive Care Unit of Henri Mondor University Hospital for ADHF requiring inotrope support. Inotropic support was added when low CO (cardiac index less than 2.2 L/minute per meter square) or impaired left ventricular (LV) ejection fraction (< 40%) was associated with low blood pressure (BP) (systolic BP < 90 mmHg or a drop in mean BP of more than 30 mmHg) or when persistent signs of low organ perfusion or oligoanuria (< 0.5 mL/kg per hour) were reported. Patients were excluded when a palliative care decision was taken, when superior vena cava catheterization failed or when associated hypovolaemic or septic shock was suspected. Patients admitted for acute myocardial infarction were excluded.

Of the 60 patients included, six had no heart failure history or known LV dysfunction and 27 (45%) had ischaemic cardiomyopathy. Admission LV ejection fraction and cardiac index averaged 25 ± 7% and 1.9 ± 0.7 L/minute per meter square (confidence interval 1.1–2.7), respectively. Current medications before hospitalization included beta-blocker therapy (70%), angiotensin-converting enzyme inhibitors (79%) and oral loop diuretics (84%). The study was approved by our local (Henry-Mondor Hospital) ethics committee and all patients gave informed consent to participate.

Catecholamine and diuretic infusion

Patients admitted for ADHF received standard care according to our local protocol, which follows current guidelines . All drugs were systematically delivered through a central venous catheter positioned via the jugular or subclavian vein. Optimal positioning of the central venous catheter was confirmed using chest radiography. The catheter had to be at the junction between the superior vena cava and the right atrium. The radiological landmark used to determine the appropriate position was the carina. Intravenous loop diuretics and vasodilator drugs were first recommended before inotropic support when the mean BP was more than 65 mmHg. In patients with low BP and clinical signs of low organ perfusion or with persistent oligoanuria, dobutamine infusion was recommended at a starting rate of 5 μg/kg per minute. Dobutamine was increased by 2.5 μg/kg per minute to a maximum of 20 μg/kg per minute, according to the physician’s clinical judgement. In patients with severe haemodynamic instability or under beta-blocker therapy at admission, dobutamine was started at the rate of 20 μg/kg per minute. Norepinephrine was given if the mean BP remained less than 65 mmHg. Intravenous loop diuretics were systematically delivered with a starting dose adjusted to the severity of volume overload, renal function and previous diuretic oral dose used. The diuretic rate was titrated to obtain euvolaemia assessed using clinical (right ventricular congestion reduction) and echocardiography data (vena cava dilatation and LV pressure reduction). Dobutamine infusion was decreased progressively (0.1 μg/kg per minute per hour) when euvolaemia and haemodynamic stability were reached. No patient needed mechanical ventilation support and, when required, oxygen was delivered through an oxygen mask to obtain an arterial oxygen saturation greater or equal to 95%.

Transthoracic echocardiography

Transthoracic echocardiography was systematically performed at admission and after 24 hours. CO was assessed using the conventional pulsed Doppler method positioned at the LV outflow tract. LV volumes and ejection fraction were quantified using Simpson’s biplane method. Early and late diastolic filling flows were assessed using conventional pulsed Doppler. Right ventricular function was quantified by the amplitude of the tricuspid annular plane systolic excursion (TAPSE). The inferior vena cava diameter was assessed from the subcostal view by M mode during the expiration period .

Haemodynamic monitoring

Invasive BP monitoring was performed using a radial arterial catheter. BP and urine output were monitored every hour during the first 12 hours and every 3 hours thereafter. ScvO ₂ assessment was performed by a standard gas analyser using a blood sample taken from the central venous catheter. ScvO ₂ was assessed before and every 12 hours after inotropic support was initiated.

Clinical outcomes

The primary endpoint was defined by the occurrence of major adverse cardiac events (MACE) that included death, heart transplantation or cardiac assistance during the hospitalization period. Heart transplantation and cardiac assistance were indicated in eligible patients (age less than 65 years without severe comorbidity) with a maximal catecholamine dose (dobutamine 20 μg/kg per minute) despite an optimal volume load or a maximal diuretic dose (furosemide 1 g/24 hours); the decision was made by an expert committee that included cardiac surgeons and cardiologists not involved in the study. The French Transplantation Agency (Agence de la Biomédecine) guidelines do not include ScvO ₂ as a criterion for patient referral for heart transplantation or cardiac assistance.

Statistical analysis

Normally distributed continuous variables were expressed as mean ± standard deviation and nominal variables as percentages. Comparisons between patients with and without MACE were done using Student’s t test or variance analysis for continuous variables and the chi-square test for dichotomous values. Paired analysis was performed for repeated values. Multivariable analysis by stepwise regression was used to identify the independent predictor of outcome. For the first step, all variables with p < 0.1 were included in the model. Statistical difference was considered as significant when p < 0.05.