An ECG showed no acute changes, and results of CT of the head were normal. A diagnosis of neutropenic septic shock secondary to hospital-acquired pneumonia was made. Broad-spectrum antimicrobials (teicoplanin, ceftazidime, amphotericin, and foscarnet) were immediately commenced, and the patient was resuscitated with sodium bicarbonate (8.4%) colloid, crystalloid, and blood products (totaling 6.5 L). Despite this, the patient remained hypotensive (80/40 mm Hg), tachycardic (110 beats/min), and anuric. Treatment was commenced with norepinephrine, and continuous venovenous hemofiltration was initiated. Over 4 h, norepinephrine doses were steadily increased to a maximum of 11 ^g/min, maintaining a MAP of > 60 mm Hg (cardiac output, 3.9 L/min; cardiac index, 2 L/min/m2; SV, 49 mL). so
Over the next 6 h, the patient again became progressively hypotensive, with a corresponding decline in cardiac output (MAP, 45 mm Hg; cardiac output, < 2 L/min; cardiac index, < 1 L/min/m2; SV, 55 mL) and required a further 3 L of colloid. Norepinephrine was replaced with escalating epinephrine infusion, reaching a maximum of 20 |j,g/min at 6 h. MAP and cardiac output remained poor (50 mm Hg; cardiac output, < 2 L/min; cardiac index, < 1 L/min/m2; SV, 40 mL). Echocardiography showed a dilated right ventricle with normal left ventricular dimensions. Pulmonary embolism was excluded by CT pulmonary angiography. In view of the low cardiac output and poor response to epinephrine and colloid, high-dose GIK was commenced (30% glucose with 50 units of actrapid insulin and 80 mmol of potassium at 1.5 mL/kg/h). This was associated with a significant rise in cardiac output, with a corresponding increase in MAP (Fig 2; Table 2).
Epinephrine was weaned to cessation within 8 h of commencing GIK. Acid-base balance had normalized, but the patient remained oliguric. The GIK infusion was stopped within 18 h of initiation. The patient remained hemodynamically stable and maintained a cardiac output from 5 to 6 L/min, with a MAP of > 70 mm Hg for the following 8 days. Glucose levels were maintained from 6 to 8 mmol/L throughout hospital admission, including the period of the GIK infusion. An abdominal ultrasound scan showed partial obstruction of the portosystemic shunt. Shunt revision was planned for the following day. However, recurrent hypotension with a high-output circulation (cardiac output, 10 L/min; cardiac index, 5.2 L/min/m2; MAP, 50 mm Hg; SV, 80 mL) led to readministration of norepinephrine. Resistant thrombocytopenia (platelets, 4 X 109/L) and markedly deranged coagulation (APTT, 50 s; TT, 18 s; PT, 30 s) resulted in pulmonary and GI hemorrhage with respiratory failure (Pao2, 45 mm Hg; Paco2, 53 mm Hg; fraction of inspired oxygen, 60%; respiratory rate, 40 breaths/min; tidal volume, 200 to 300 mL; pressure support, 20 cm H2O; positive end-expiratory pressure, 10 cm H2O). With full family consent, treatment was withdrawn and the patient died. In both cases, activated protein C, corticosteroids, vasopressin, or other adjuvant therapies for sepsis were not administered.
Figure 2. GIK infusion in patient 2: sustained improvement in cardiac output and decrease in epinephrine dose following introduction of high-dose GIK. See Figure 1 legend for definition of abbreviation.
Table 2—Hemodynamics and Glucose Levels Before and After GIK Administration in Patient 2
|Variables||Before GIK (Immediate)||After GIK (30 min)||GIK Stopped(18 h)|
|MAP, mm Hg||52||70||65|
|Heart rate, beats/min||115||110||90|
|CVP, mm Hg||18||19||17|
|Cardiac output, L/min||2.5||5.5||5.0|
|Cardiac index, L/min/m2||1.3||2.8||2.65|
|Epinephrine dose, |xg/min||20||20||0|
|Mean glucose level, mmol/L||6.8||7.2||7.8|