AUTOMATIC VENTILATION CONTROL TO MINIMIZE VENTILATOR INDUCED LUNG INJURY

A. V. Pino§,*, A. Roncally*, F. Lima*, F. A. Bozza ‡, J. Salluh ‡, F. Ascoli †, J. Neves †, F. C. Jandre *, A. Giannella-Neto *

* Biomedical Engineering Program/COPPE, Federal University of Rio de Janeiro, Brazil,
§ Biomedical Engineering Group, Catholic University of Pelotas, Brasil
† Veterinary Pharmacology and Physiology Department, Fluminense Federal University, Rio de Janeiro, Brasil
‡ ICU of Clementino Fraga Filho Hospital, UFRJ, Rio de Janeiro, Brasil
pino@atlas.ucpel.tche.br

Abstract: Setting mechanical ventilation parameters to optimize alveolar recruitment and prevent alveolar overdistension (AO), has been receiving much attention. A swine model of acute respiratory distress syndrome (ARDS) was used to test a closed-loop ventilation controller designed to find the positive endexpiratory pressure (PEEP) for minimum respiratory system elastance, and the tidal volume that prevents AO. The automatic ventilation was compared to a conventional protective approach. PEEP values obtained with this controller were similar to those obtained with manual titration. The AO appeared to be lower in automatic than in conventional ventilation. The controller was stable and the results recommend its use for ARDS therapy.

Introduction

Despite recent advances in treatment, the mortality in the acute respiratory distress syndrome (ARDS) remains high. ARDS patients need mechanical ventilation support, but the best adjustment for tidal volume (VT) and positive end expiratory pressure (PEEP) are not consensual [1]. Nowadays, physicians try to set these two parameters in order to avoid mechanical stress in lung tissues.

The use of low VT (6-7 ml/kg) strategy has been earning credibility since the ARDSNet group showed its efficiency as a lung protective therapy [2] to reduce the mortality associated with ARDS. The use of low VT is thought to reduce alveolar overdistention (AO) and pul- monary inflammation.

The PEEP can be set to open and the small airways and to prevent it to close during respiratory cycle. In order to avoid AO, the lower inflection point, drawn from the pressure-volume curve, has been proposed as a good compromise to adjust PEEP [3,4].

This work presents a closed-loop controller to act over VT and PEEP, in order to open the lung and to avoid AO in a porcine model with oleic acid-induced lung injury.

Methods

The protocol was approved by the Ethical Commission for Assessment of Animal Use in Research (CEUA-FIOCRUZ). Twelve female piglets, (15-20 kg), were anesthetized with cis-ketamin (10 mg/kg/h), blocked with pancuronium bromide (2 mg/kg/h) and ventilated through an endotracheal tube of 6.0 or 7.0 mm. An Amadeus (Hamilton, Medical Suisse) ventilator was used with an additional electronic board to allow external ventilator control by a serial port. ARDS was obtained through oleic acid injections. Two groups of six piglets each were studied. The first group, named control, was ventilated with low VT (6-7 ml/kg), inspired oxygen fraction (FIO2) of 1.0, and respiratory frequency of 20-30 cpm. PEEP was choose to minimize the volume independent elastance (E1, in equation 1) during a step descendent titration PEEP procedure.

P( t ) = R·F( t ) + (E1 + E2·V( t ) ) ·V( t ) + PEEP (1)

were E1 is the volume independent elastance, E2 is the volume dependent elastance, R is resistance, P is airway opening pressure, F is flow and V is volume.

In the second group, named automatically ventilated group, PEEP was chosen to minimize E1 using a unidirectional gradient descent algorithm, modified from Jandre and coworkers [5], using a linear approximation of E1 at each PEEP:

E1 (PEEP) = E1o + γ·PEEP (2)

The VT was set according to an index of alveolar overdistention (%E2, described by Kano and coworkers [4]), as the VT that kept %E2 equal to 30%:

VT = ( 0.3·E1 ) / ( 0.7·E2 ) (3)

This choice of VT intended to bear the highest VT that avoids AO. VT was bounded at a minimum of 6 mL/kg and at a maximum of 9 mL/kg, assumed as safety values.

Both the parameter set R, E1 and E2 (from Equation 1) and g (Equation 2) were obtained by linear recursive least squares algorithms, with a time constant of 20 s [6].

The control process started with a recruitment maneuver, consisting in using a continuous positive airway pressure of 30 cmH2O during 25 s. Then, the best PEEP was set using the gradient descent method. After the PEEP adjustment the VT controller was turned on. If it was not possible to avoid AO within the bounds, the VT controller was turned off and the recruitment maneuver was restarted.

Results

Figure 1 shows the results of %E2 and Figure 2 shows the PEEP results of the two animal groups.

Figure 1:Alveolar overdistension index (%E2) over 4 hours after ARDS onset.
Figure 1:Alveolar overdistension index (%E2) over 4 hours after ARDS onset.
Figure 2: Positive end expiratory pressure (PEEP) over 4 hours after ARDS onset.
Figure 2: Positive end expiratory pressure (PEEP) over 4 hours after ARDS onset.

Discussion

The mean PEEP obtained in the two groups was similar, but the variability in the control group was higher. The %E2>30% was present in 3 out of 6 piglets in the control group (Figure 1), and only in 1 out of 6 piglets in the automatically ventilated group. The piglet with AO in the automatically ventilated group presented airways secretion, which could have contributed for non adequate estimates of the mechanical parameters includ- ing E1, E2 and %E2. The gradient descent method seemed to result in a more stable lung. The %E2 index was lower in the automatically ventilated group al- though a higher VT was allowed for this group.

Conclusion

Even at low VT, AO may appear, as observed in 3 of 6 piglets in the control group. The control algorithms were stable and maintained %E2 below 30%. These results justify the use of this controller in order to main- tain VT sufficiently low for preventing AO.

Acknowledgments

We would like to acknowledge the partial financial support of the brazilian agencies CNPq and FAPERJ, and the technical advise of Luiz Costa and Luciano Kagami.

References

  1. Esteban A., Anzueto A., Alia I., Gordo F., Apezteguia C., Palizas F., Cide D., Goldwaser R., Soto L., Bugedo G., Rodrigo C., Pimentel J., Raimondi G., Tobin M. J., "How is mechanical ventilation employed in the intensive care unit? An international utilization review", American Journal of Respiratory and Critical Care Medicine 161(5):1450-1458, 2000.
  2. ARDSNet, "Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network", The New England Journal of Medicine 342(18):1301-1308, 2000.
  3. Vieira S. R., Puybasset L., Lu Q., Richecoeur J., Cluzel P., Coriat P., Rouby J. J., "A scanographic assessment of pulmonary morphology in acute lung injury. Significance of the lower inflection point detected on the lung pressure-volume curve", American Journal of Respiratory and Critical Care Medicine 159(5 (Pt 1)):1612-1623, 1999.
  4. Kano S., Lanteri C. J., Duncan A. W., Sly P. D., "Influence of nonlinearities on estimates of respiratory mechanics using multilinear regression analysis", Journal of Applied Physiology 77(3):1185-1197, 1994.
  5. Jandre F. C., Pino A. V., Lacorte I., Soares J. H. N., Giannella-Neto A., "A closed-loop mechanical ventilation controller with explicit objective functions", IEEE Transactions on Biomedical Engineering 51(5):823-831, 2004.
  6. Lauzon A. M., Bates J. H., "Estimation of timevarying respiratory mechanical parameters by recursive least squares", Journal of Applied Physiology 71(3):1159-1165, 1991.