Accuracy of tidal volume delivery by five different models of large-animal ventilators

To determine the accuracy of tidal volume (V_T) delivery among 5 different models of large-animal ventilators when tested at various settings for V_T delivery, peak inspiratory flow (PIF) rate, and fresh gas flow (FGF) rate.

4 different models of pneumatically powered ventilators and 1 electrically powered piston-driven ventilator.

After a leak flow check, each ventilator was tested 10 times for each experimental setting combination of 5 levels of preset V_T, 3 PIF rates, and 4 FGF rates. A thermal mass flow and volume meter was used as the gold-standard method to measure delivered V_T. In addition, circuit systems of rubber versus polyvinyl chloride breathing hoses were evaluated with the piston-driven ventilator. Differences between preset and delivered V_T (volume error [ΔV_T]) were calculated as a percentage of preset V_T, and ANOVA was used to compare results across devices. Pearson correlation coefficient analyses and the coefficient of determination (r²) were used to assess potential associations between the ΔV_T and the preset V_T, PIF rate, and FGF rate.

For each combination of experimental settings, ventilators had ΔV_T values that ranged from 1.2% to 22.2%. Mean ± SD ΔV_T was 4.8 ± 2.5% for the piston-driven ventilator, compared with 6.6 ± 3.2%, 10.6 ± 2.9%, 13.8 ± 2.97%, and 15.2 ± 2.6% for the 4 pneumatic ventilators. The ΔV_T increased with higher PIF rates (r² = 0.69), decreased with higher FGF rates (r² = 0.62), and decreased with higher preset V_T (r² = 0.58).

Results indicated that the tested ventilators all had ΔV_T but that the extent of each of ΔV_T varied among ventilators. Close monitoring of delivered V_T with external flow and volume meters is warranted, particularly when pneumatic ventilators are used or when very precise V_T delivery is required.