Asynchronous Independent Intermittent Positive Pressure Ventilation as a Solution to Refractory Hypoxemia during Chest Surgery

A 63-yr-old man was diagnosed as suffering from T5–T6 vertebrae osteomyelitis and paravertebral abscess. His past medical history included chronic renal failure, non–insulin-dependent diabetes mellitus, hypertension, hypothyroidism, and epilepsy. The patient underwent an uneventful emergent abscess drainage and posterior spine fusion in the prone position. On the third postoperative day, the patient required a second emergent operation due to worsening of paraparesis. Anterior thoracic spine fusion via  right thoracotomy was indicated. Blood tests were significant for abnormal results of renal function tests. Arterial blood gases on the day before were as follows (fraction of inspired oxygen [Fio²], 0.21): pH, 7.37; partial pressure of carbon dioxide (Pco²), 39 mmHg; and partial pressure of oxygen (Po²), 84 mmHg. A chest radiograph prior to surgery showed atelectasis of the left lower lobe (fig. 1). Anesthesia was induced with midazolam, etomidate, and fentanyl intravenously. Vecuronium bromide was administered to facilitate tracheal intubation. The trachea was intubated with a 41-French, left-sided, double-lumen tube (Broncho-cath™; Mallinckrodt Medical, Athione, Ireland). Tube positioning was confirmed by auscultation and fiberoptic bronchoscopy in the supine as well as in the left lateral decubitus position. Patient monitoring included (in addition to standard monitoring) radial artery catheter for invasive blood pressure measurement and a left internal jugular vein catheter. Anesthesia was maintained with isoflurane and 50% nitrous oxide in oxygen supplemented by intermittent boluses of fentanyl and vecuronium. The lungs were ventilated with a tidal volume of 10 ml/kg; respiratory rate was adjusted to maintain normocapnia and PEEP of 4 cm H²O. Arterial blood gas values and ventilation parameters are shown in table 1.

Several minutes before nondependent right lung deflation, oxygen concentration was raised to 100%. Once the chest was opened and the right lung deflated, a significant decrease in the arterial saturation (from 98% to 63%) was noted. Oxygenation did not improve significantly either by 10 cm H²O CPAP applied to the nondependent lung initially or by 10 cm H²O PEEP application to the dependent lung later. Fiberoptic bronchoscopy confirmed proper positioning of the double-lumen tube without mucous plugging or increased secretions. Restoration of two-lung ventilation improved oxygen saturation, but surgery could not continue because the right lung obstructed the surgical field. High-frequency jet ventilation was considered, but the ventilator was unavailable.

At this stage, differential lung ventilation was initiated as follows (table 1): the dependent lung was ventilated by the anesthesia machine with a tidal volume of 6 ml/kg, a respiratory rate of 14 breaths/min, a PEEP of 10 cm H²O, and an Fio²of 1.0. The nondependent lung was ventilated by a volume control, battery-powered, servo controlled, transport ventilator (Flight III; Flight Medical, Lod, Israel) with a tidal volume of 1 ml/kg, a respiratory rate of 20 breaths/min, a PEEP of 0 cm H²O, and an Fio²of 1.0. Peak inspiratory pressures were 35 and 25 cm H²O, respectively. Under this ventilation mode, oxygenation increased dramatically to 100% oxygen saturation (table 1). The right, nondependent lung volume was similar to that seen without ventilation; respiratory inflations resulted in small volume changes that did not interfere with the surgical exposure. The operation continued uneventfully, and the patient was extubated in the operating room at the end of surgery. The patient was discharged from the intensive care unit 24 h later and transferred back to the ward.

One-lung ventilation is routinely used during thoracic surgery. It provides good exposure and adequate surgical conditions. However, OLV creates an obligatory right-to-left transpulmonary shunt through the nondependent, nonventilated lung and increases alveolar-to-arterial oxygen tension difference, so hypoxemia may develop. Blood flow to the nondependent lung is usually reduced by gravity, lung collapse, surgeon manipulation, and by active hypoxic pulmonary vasoconstriction, thus decreasing the shunt. 3Hence, conventional ventilation of the dependent lung with 100% oxygen is usually associated with acceptable arterial oxygen tension (Pao²) values. If hypoxemia does occur, optimization of oxygenation can be obtained in several ways. 4,5CPAP with 100% oxygen to the collapsed lung is the most effective means for correction of hypoxemia during OLV. 10PEEP applied to the ventilated lung can recruit collapsed alveoli and improve oxygenation. However, PEEP increases the dependent lung alveolar pressure and pulmonary vascular resistance, thus diverting blood to the nondependent, nonventilated lung and, in turn, worsening the shunt and hypoxemia. 1The result of PEEP application to the dependent lung is therefore unpredictable, usually favorable only if the dependent lung is basically ill. 11 

Baraka 7described a method of differential intermittent positive pressure ventilation by partial occlusion of the adapter limb to the nondependent lung while maintaining unrestricted ventilation of the dependent lung. Yamamura et al.  8invented a complicated, large, and heavy device, which enables the diversion of different proportions of the tidal volume to the nondependent lung, thus ventilating it differentially. Both methods lack the ability to monitor and control precisely volumes and pressures.

In our patient, hypoxemia developed shortly after instituting OLV. The common practices of treating hypoxemia during OLV were taken (100% oxygen, reconfirmation of tube patency and proper position, CPAP to the nondependent lung and then PEEP to the dependent lung) but failed to improve oxygenation. In this report, we describe a simple technique for differential lung ventilation that improved severe hypoxemia related to OLV. Our technique has three major advantages. First is the use of simple equipment, which is highly accessible. Second is the ability to determine and monitor precisely ventilation parameters. The third advantage is that using the anesthesia machine to ventilate the dependent lung keeps the ability to maintain inhaled anesthesia (considering the negative effect on the hypoxic pulmonary vasoconstriction). The small tidal volume applied to the nondependent lung yielded small changes in the lung volume, which did not interfere with the surgical exposure. Peak inspiratory pressure of the nondependent lung was high relative to the small tidal volume (25 cm H²O). This high peak inspiratory pressure can be attributed to low compliance of a lung that was deflated for a while, small diameter of the tracheal lumen, direct surgical pressure, or any combination of these three explanations. It seems that gas trapping was not evident since lung volume was small and did not increase with time. In addition, a slightly increased respiratory rate (20 breaths/min) should allow sufficient time for small tidal volume expiration (auto PEEP could not be measured with our equipment). Blood pressure did not change while the lungs were ventilated differentially; therefore, there was no need for synchronization between the ventilators. 11 

In conclusion, this case demonstrates the use of asynchronous differential intermittent positive pressure ventilation as a simple and useful way to improve oxygenation during OLV.