An abnormal capnography trace due to air embolism in the lateral position

Abstract

Venous air embolism occurs when air is entrained into the venous system and travels to the right heart and pulmonary circulation, and commonly occurs as a complication in laparoscopic, neurosurgical and cardiac surgeries. We present a case of abnormal end-tidal carbon dioxide capnography tracing in the lateral position in a laparoscopic major liver procedure and discuss the potential novel use of this as a red flag in aiding the medical practitioner to diagnose air embolism.

Background

With a shift in surgical techniques favouring minimally invasive procedures, cases involving the use of pneumoperitoneum such as laparoscopic and robotic surgeries are expected to be the approach of choice for many surgeries. Unfortunately, while there are many merits of minimally invasive surgeries, one of the most feared and potentially devastating condition remains vascular air embolism (VAE). The ‘gold standard’ for detection of VAE is transeophageal echocardiography as it is highly sensitive but its use is limited as it requires expertise and is expensive. We discuss an unusual capnography tracing in a patient who underwent laparoscopic hepatectomy complicated by suspected air embolism and explore the possible reasons for it.

Case presentation

A 71-year-old man (weight 83 kg, height 1.72 m, body mass index (BMI) 28.1) was admitted electively for a laparoscopic resection of hepatic segments 6 and 7 with exposure of right hepatic vein. He had a background of Childs A liver cirrhosis from hepatitis B complicated by hepatocellular carcinoma 1.4x1.1 cm in segment 6, diabetes mellitus, chronic renal impairment, dyslipidemia and hypertension. The tumour was incidentally picked up during routine ultrasound abdomen surveillance for his hepatitis B and he was otherwise asymptomatic with no features of decompensated liver cirrhosis. He had a good effort tolerance with a preoperative echocardiograph showing a left ventricular ejection fraction of 66% and a negative stress echocardiograph.

Prior to induction of anaesthesia, an arterial and internal jugular central venous line were inserted. She was induced with propofol, rocuronium and fentanyl, maintained on an oxygen/air/desflurane mixture and placed in the left lateral position for optimal surgical access and analgesia was administered via targeted controlled infusion of remifentanil infusion of between 0.5 and 2 ng/mL. A fluid restrictive therapy was utilised. Urine output was estimated to be about 100 mL over 5.5 hours since induction. Pneumoperitoneum was induced with carbon dioxide (CO₂) and pressures were kept at 12 mm Hg.

3.5 hours into the procedure, the capnograph was noted to have changed from the normal plateau to a saw-tooth pattern. Shortly after, the patient was noted to be acutely getting more hypotensive (blood pressure (BP) 115–120/80–85 to 65–80/50–70 mm Hg, mean arterial pressure (MAP) 90–95 to 50–65 mm Hg), tachycardic (heart rate 88–108 bpm), with a corresponding drop in the end-tidal carbon dioxide (ETCO₂) from 40→16 mm Hg and pulse oximeter reading from 100% on FiO₂ 0.36 to 59% on FiO₂ 0.88 (figure 1A). Surgeons were informed of the patient’s unstable hemodynamics and disclosed that they had just transected a large vessel with ~200 mL blood loss over 15 min. A presumptive diagnosis of air embolism was made and pneumoperitoneum was deflated, patient placed on 100% O₂ and blood pressure supported by fluid boluses and vasopressors (phenylephrine titrated boluses of 3.3 mg and noradrenaline of 0.01–0.05 μg/kg/min). Once his haemodynamics were stabilised, 40 mL of blood with no air or froth was aspirated from the central catheter. Surgery thereafter resumed but there was a recurrence of the signs for another two episodes with supportive therapy being administered. The rest of the surgery proceeded uneventfully with an estimated blood loss of 2 L and surgery was concluded 2.5 hours later. An arterial blood gas was performed at the end of surgery showing respiratory acidosis (pH 7.25, pCO₂ 51 mm Hg, pO₂ 115 mm Hg, base excess (BE) –5 mmol/L, HCO₃ 22 mmol/L on FiO₂ 0.35), and the patient was extubated and sent to the intermediate care unit.

Figure 1.

(A) Vital sign and capnogram trace with the sawtooth pattern. (B) Blood flow is restored as P_A decreases; this occurs at different P_A. (C) Oscillating end-tidal carbon dioxide waveform possibly resulting from intermittent alveolar perfusion and hence differential rates of carbon dioxide diffusion during expiration.

Outcome and follow-up

The patient was extubated and sent to the intermediate care unit. Unfortunately in the postoperative period, the patient developed a hospital acquired pneumonia and fast atrial fibrillation and had a delayed discharge home about 2 weeks after the surgery.

Discussion

Venous air embolism is a potentially catastrophic complication which has been well described in association with pneumoperitoneum. Porcine studies have consistently demonstrated high incidences of air embolism for laparoscopic hepatic resections performed under CO₂ pneumoperitoneum with rates as high as 66%–100% with up to 50% of these associated with cardiorespiratory compromise.¹ In particular, for patients with underlying liver cirrhosis, the presence of abnormal arteriovenous communications increases their risk of paradoxical emboli which may bypass the pulmonary circulation and enter directly into the systemic circulation.² It is henceforth important for both surgeons and anaesthesiologists to be aware of this condition, recognise the early features and know how to manage it.

Laparoscopy is preferred in hepatectomies as it is associated with reduced blood loss, is more cosmetically pleasing and has reduced postoperative pain. However, the incidence of gas embolism is high in these surgeries due to the highly vascular liver bed, low central venous pressure (CVP) and prolonged surgical time. Compared with traditional open hepatectomies, VAE also occurs more frequently in CO₂ pneumoperitoneum as CO₂ is highly soluble and quickly absorbed into the systemic circulation. Those at risk include major liver resections (right or left hemihepatectomies or posterior sectionectomies) where the major hepatic veins are exposed to the atmosphere, and any air entrained will be transported directly into the inferior vena cava. In contrast, for minor liver resections, there is less exposure and dissection of the major hepatic vessels with a corresponding lower risk of air embolism. Air embolism may occur during all phases of dissection but appears to have a predisposition for two periods: opening of the liver capsule, and deep parenchymal transection (>80%).³

The use of end-tidal capnography has been described since 1975 and has been shown to be a reliable non-invasive means to detect pulmonary embolism. The classical capnography trace represents an exponential decrease in ETCO₂ levels. This occurs due to the dilution of CO₂ from ventilated but non/poorly perfused alveoli. Interestingly, however in our case, we noted an irregular saw-tooth ETCO₂ waveform. We postulate that this could be due to the differential perfusion and ventilation pattern of the lungs, which may account for the different capnography trace.

In normal lungs, most of the alveoli are of West zone 2. Blood flow through a pulmonary capillary is determined by P_a>P_A>P_v, where P_a is pulmonary arteriole pressure, P_A is alveolus pressure and P_v is pulmonary venule pressure. As microemboli are carried in the blood stream, they distribute preferentially to the dependent lung following perfusion rather than ventilation. During a mechanically ventilated inspiration breath, blood flow occlusion by microemboli is exacerbated by extrinsic compression of the pulmonary vasculature at end-inspiration (high P_A) with a dynamic increase in West zone 1 alveoli. During expiration, as P_A decreases, West zone 1 alveoli evolve to zone 2. Depending on the size of the microemboli, blood flow is reinstituted at various phases of expiration depending on P_A. The irregular return of West zone 1 to 2, and 2 to 3, explains the saw-tooth appearance on capnography during expiratory phases (figure 1B and C). An irregular capnography waveform may hence be used to support the early diagnosis of a venous air embolism. However, further studies will need to be performed to confirm our hypotheses.

For air to enter into the vascular system, either an intravascular negative pressure or a high positive extrinsic pressure must be present. Pneumoperitoneal pressures of <12 mm Hg are believed to be sufficient for surgical access with low rates of gas embolism, although more trials are needed to confirm this.⁴ Low CVP have been used as a surrogate marker of cardiac preload, and hence the degree of hepatic vein congestion in order to reduce intraoperative bleeding in liver surgeries and thereby improving outcomes.⁵ Unfortunately, a low CVP and relative hypovolemia will exacerbate any incidences of air embolism.

The choice of anaesthetic maintenance appears to play a part in air embolism. A randomised controlled trial on the use of sevoflurane in liver surgeries found that sevoflurane is associated with a longer mean embolism duration (51 vs 34 s, p<0.05), worse pulmonary blood gas exchange (PO₂/FiO₂ 451 vs 504, p<0.05) and haemodynamics (MAP 73 vs 82 mm Hg, p=0.01) than propofol.⁴ The pulmonary circulation has long been recognised to be an important physiological filter for air emboli and its filtering capacity is affected by the use of anaesthetics—the pulmonary system is able to tolerate a much lower volume of air in anaesthesia with halothane compared with fentanyl or ketamine. The mechanism for this phenomenon remains unclear but is suspected to be mediated at the arteriolar, shunt vessels or capillary level.⁶

The diagnosis of air embolism remains clinical but is highly suggestive in the presence of acute hypotension, tachycardia, hypoxia, drop in ETCO₂ with a positive air aspiration from the central venous catheter (CVC) and development of a mill-wheel murmur. Unfortunately, in the majority of cases, a positive air aspiration is not present, and haemodynamic instability, hypoxia and arrthymias occur late. Treatment is supportive. Fortunately, the effects of CO₂ embolism are usually short lived as CO₂ has high blood solubility. However, significant haemodynamic instability and cardiovascular collapse can happen if rapid entrainment of large volumes of air (exceeding 3–5 mL/kg) occurs as this will cause pulmonary vasoconstriction, increase pulmonary pressures, leading to acute right heart failure.

Learning points.

• The presence of an early saw-tooth capnograph preceding hypotension and hypoxia should raise the suspicion of an air embolism.

• This is in addition to the classical features of air embolism such as acute hypotension, tachycardia, hypoxia, drop in end-tidal carbon dioxide with a positive air aspiration from the CVC and development of a mill-wheel murmur.

• Patients undergoing hepatectomies are at increased risk of developing air embolism due to use of laparoscopic gases, exposure of major hepatic vessels to the atmosphere and low targeted central venous pressures.