Persistent preload defect in severe sepsis despite fluid loading: A longitudinal echocardiographic study in patients with septic shock.

STUDY OBJECTIVE: To investigate the rate of recovery from septic shock in patients with suspected left ventricular (LV) preload deficiency and LV systolic dysfunction.

DESIGN: A monitoring period was defined by the need for inotropic/vasopressor support, and LV function was assessed daily during this period by bedside two-dimensional echocardiography (2D-ECHO).

SETTING: University hospital ICU.

PATIENTS: During a 5-year period, 90 patients with an episode of septic shock (60% with gram-positive bacteria as the causative agent) were consecutively enrolled in the study (mean age, 55 +/- 18 years). Standard volume resuscitation combined with inotropic/vasopressor support was used to maintain systolic arterial pressure > 90 mm Hg. All patients received mechanical ventilation because of associated respiratory failure. The average duration of hemodynamic support was 4.4 +/- 1.6 days. Thirty-four patients were weaned from hemodynamic support during the monitoring period and ultimately recovered (group I). Twenty-eight patients died from refractory circulatory failure during the monitoring period, and 28 died later from ARDS or multiple organ dysfunction syndrome, leading to a 62% overall mortality rate (group II).

METHODS: Daily bedside LV volumes and ejection fraction (LVEF) were recorded using 2D-ECHO. Data obtained at the start (day 1 and day 2) and end of the monitoring period (day n) were compared.

RESULTS: LV end-diastolic volume was within the normal range of our laboratory values in all patients, but was initially smaller in group II than in group I, and remained so despite fluid loading. LVEF was significantly depressed in all patients, resulting in severe reduction in LV stroke volume (LVSV), which was initially more marked in group I. In group I patients, LVEF significantly improved during the monitoring period, resulting in an increase in LVSV.

CONCLUSION: 2D-ECHO changes during hemodynamic support in 90 septic patients confirmed defective LV preload with a propensity to worsen despite fluid loading in nonsurvivors (62% in the present study). Our results are also in agreement with previous studies reporting depressed LV systolic function at the initial phase of septic shock. Since LV dysfunction was more marked in patients who recovered, we suggest that the exact significance of this finding should be reevaluated.