2006 OPEN FORUM Abstracts
PERFORMANCE OF MECHANICAL VENTILATORS IN THE SETTING OF PATIENT EFFORT: A BENCH STUDY
Juliana C Ferreira, MD
Daniel W Chipman RRT, Robert M Kacmarek PhD RRT. Massachusetts General Hospital
and Harvard Medical School, Boston MA.
Background: Mid-range mechanical
ventilators have become increasingly available from many manufacturers and are
used in many hospitals. However, the performance characteristics of these
ventilators are unknown, especially in the presence of impaired respiratory
mechanics and high patient effort. We used a mechanical lung model to compare
the performance of 10 mid-range mechanical ventilators with a standard ICU
ventilator.
Methods: The PB760 (Puritan Bennett), Esprit (Respironics),
eVent (eVent Medical), Ivent (VersaMed),
HT50 (Newport Medical), LTV 1000 (Pulmonetics), Raphael (Hamilton), Savina
(Drager Medical), Servo S (Maquet) and Vela (Viasys) ventilators were compared
to the PB840 (Puritan Bennett). The mechanical lung model used was the ASL5000
simulator (IngMar), a computerized mechanical simulator that can generate
patient efforts of various magnitudes, and allows the setting of system
compliance and resistance to mimic normal, obstructive and restrictive
conditions. Two levels of effort were tested, peak flows of 30 and 60 L/min.
The ventilators were set in pressure assist/control mode, with a inspiratory
pressure 10 cm H2O, PEEP 5 cm H2O, and sensitivity set as
sensitive as possible avoiding auto-cycling. The variables analyzed were
triggering delay in ms (TD), triggering work in mJ (WOBt), maximum pressure
drop during triggering in cmH2O (Pmin), and ventilator inspiratory
work in mJ/L (WOBi,v). Differences among ventilators greater than 10% were
considered clinically significant.
Results: Performance varied widely
among the ventilators, and had a similar pattern in normal, obstructive and
restrictive conditions. Higher patient effort was associated with poorer
performance for most variables with all ventilators. TD was <100ms in most
conditions for the Esprit, PB760, Servo S and Vela, and was >150ms in most
conditions only for the HT 50. WOBt was similar to the PB840 (0.64mJ with
normal mechanics and 60L/min demand) for the PB760, Raphael and Servo S,
approximately 2 times greater for the LTV 1000 and eVent, and 5 times greater
for the HT50, Ivent, and Savina. WOBi,v was similar to the PB840 (1064mJ/L with
normal mechanics and 60L/min demand) for the PB760, Esprit, Ivent, LTV 1000 and
Servo S, was 20% lower in a few conditions for the Raphael and Vela, and was
consistently lower for the eVent, HT50 and Savina. Pmin was equivalent to that
of PB840 (-1.33 cm H2O with normal mechanics and 60 L/min demand)
for the Vela and Esprit, less than 50% greater for the PB760, LTV 1000, Savina
and Servo S, and two to three times greater for the eVent, Ivent, HT50 and
Raphael.
Conclusions: Most ventilators performed slightly poorer than
the PB840 in this model of high patient effort. The Esprit and the PB760 were
the best performers of the group. The Newport, Ivent and eVent demonstrated a
wide variability among breaths. These results indicate that mid-range
ventilators may not perform as well as ICU ventilators and special attention
should be given to patient sedation and synchrony during use to minimize the
impact of high patient effort on ventilator performance.