Desaturation during one-lung ventilation
66-year-old woman, 160 cm and 78 kg, former smoker with FEV1 78% and DLCO 68% predicted, scheduled for right upper lobectomy through a thoracotomy. After induction a 35 Fr left-sided double-lumen tube is placed, and fiberoptic bronchoscopy confirms the bronchial cuff in the left main bronchus with the carina above the tracheal cuff. She is positioned in the right lateral decubitus and one-lung ventilation is started: volume-controlled, tidal volume 450 mL, PEEP 5 cmH2O, FiO2 0.5, rate 14/min. Ten minutes later SpO2 is 88%, peak airway pressure 34 cmH2O (22 cmH2O during two-lung ventilation), plateau pressure 28 cmH2O, EtCO2 49 mmHg and rising, heart rate 92/min, blood pressure 112/66 mmHg. The chest is open and the surgeon says the right lung is still partly inflating with each breath. Breath sounds are present on the left and diminished at the right apex. A radial arterial line was placed before induction.
1. List what you check first, in order, and explain the physiology that drives the desaturation.
Read the worked answer
1. Tube position after positioning. Double-lumen tubes migrate in a substantial minority of patients when they are turned and the chest is opened. Re-scope: tracheal cuff at the carina, bronchial cuff just below it in the left main bronchus. A tube drawn back into the trachea, or a bronchial cuff herniating over the carina, produces exactly this triad - high pressures, rising EtCO2, and an operative lung that still inflates.
2. Whether isolation actually exists. The surgeon can see the right lung moving. Check the bronchial cuff volume, both lumens, the Y-piece and the cap on the non-ventilated lumen. Even a small leak pressurises the operative lung and abolishes the shunt reduction you are relying on.
3. The dependent lung. Suction for secretions or blood, auscultate, and send an arterial blood gas. High peak pressures with a high plateau point to the lung, not the tube.
Physiology. One-lung ventilation creates a large right-to-left shunt through the non-ventilated lung. Hypoxic pulmonary vasoconstriction diverts blood away from it, but this is a slow, partial response - it takes 20-30 minutes to plateau and diverts at most about half of the flow. Volatile agents, acidosis, vasodilators, high mixed venous oxygen saturation and high pulmonary artery pressures all blunt it. While you investigate, turn FiO2 to 1.0: it does not fix the shunt, but it maximises the oxygen carried by the ventilated lung, and a rising shunt fraction is a physiological emergency, not a monitor problem.
2. Isolation is confirmed, the tube is correctly positioned, and SpO2 is 89% on FiO2 1.0 with a stable blood pressure. Set out your escalation, explaining why each step can backfire.
Read the worked answer
Step 1 - optimise the dependent lung. Suction, ensure the circuit has no leak, and set PEEP to the dependent lung in small increments (5 to 8 cmH2O) while watching the plateau pressure. Backfire: PEEP that overdistends the ventilated lung raises its vascular resistance and diverts blood towards the non-ventilated lung, worsening the shunt. Stop if the plateau climbs above about 25-27 cmH2O or the saturation falls.
Step 2 - recruit the operative lung. Apply CPAP 2-5 cmH2O to the non-ventilated lung if the surgeon will tolerate it. Backfire: CPAP inflates the operative lung and obscures the field; above about 5 cmH2O it defeats the purpose of the isolation.
Step 3 - intermittent two-lung ventilation. Ask the surgeon for a pause, re-inflate both lungs for 2-3 minutes, then resume. Useful but only buys time.
Step 4 - pharmacological and mechanical support. Correct anemia and hypotension, keep the cardiac output adequate, avoid excessive volatile agent, and consider adding an inotrope if mixed venous saturation is low.
Step 5 - surgical options. Ligation or clamping of the pulmonary artery to the operative lobe, or packing the lung, if the patient cannot tolerate OLV.
Step 6 - ECMO is a last resort and requires a plan before the case starts. Throughout, accept a saturation of 88-90% if the patient is stable and the lactate is not rising; 85% with hemodynamic compromise is not acceptable and OLV should be abandoned.
3. The current settings are tidal volume 450 mL with PEEP 5 cmH2O. Calculate the predicted body weight, the driving pressure, and the ventilator changes you would make, with the reasoning.
Read the worked answer
Predicted body weight (female): 45.5 + 0.91 × (160 − 152.4) = 45.5 + 6.9 ≈ 52.4 kg. Tidal volume should be 6-8 mL/kg PBW, i.e. 315-420 mL. At 450 mL the patient is being ventilated at about 8.6 mL/kg PBW, which is above the lung-protective range and is one explanation for the high airway pressures.
Driving pressure: plateau minus PEEP = 28 − 5 = 23 cmH2O. Targets during one-lung ventilation are a plateau below about 25 cmH2O and a driving pressure below about 15-16 cmH2O; both are exceeded.
Changes:
- Reduce tidal volume to 350-380 mL and accept permissive hypercapnia with a pH above 7.20, raising the rate to keep the EtCO2 near 45 mmHg.
- Keep PEEP at 5 cmH2O initially and titrate upward only if oxygenation demands it and the driving pressure allows.
- Re-measure the plateau after each change; the trend matters more than any single number.
- Re-scope after any ventilator change that alters the position of the tube or the surgical retraction.
Reasoning: high tidal volumes and high driving pressures in the ventilated lung cause volutrauma and ventilator-induced lung injury, and they also raise pulmonary vascular resistance, which pushes blood towards the non-ventilated lung and worsens the shunt. Lowering the tidal volume therefore improves both lung protection and, often, oxygenation.