Bedside Diagnosis of Pulmonary Embolism Using Electrical Impedance Tomography

Computed tomography (CT) angiography is the gold standard for diagnosing pulmonary embolism (PE).¹ However, CT imaging can have important limitations in critically ill patients, many of whom may have organ dysfunction or contraindications to receiving contrast. The transfer of these patients from intensive care units (ICUs) to a scanner can also be resource- intensive and dangerous, depending on patient stability. A recent meta-analysis found that in a total of 12,313 intrahospital transports of critically ill patients, there was a 26.2% pooled frequency of adverse events.²

Electrical impedance tomography (EIT) is a noninvasive imaging modality that can be done at the bedside to assess for regional lung perfusion defects. This case report details the use of EIT in diagnosing PE in a critically ill patient with refractory hypoxemia who was initially too unstable to transport for CT imaging. It highlights how EIT can fill an important need for bedside diagnostics for a critical diagnosis such as PE in unstable patients.

The ventilatory tools of EIT are approved by the US Food and Drug Administration (FDA). Written Health Insurance Portability and Accountability Act (HIPAA)-compliant consent to present this case was provided by the patient.

CASE DESCRIPTION

A 42-year-old man with a history of a remote abdominal gunshot wound presented for an elective ventral hernia repair. During surgery, the infraumbilical fascia was closed, but the surgical team was concerned for a tight closure, and the patient’s abdomen was left open. He was intubated and transferred to the ICU for close monitoring.

The patient had an acute desaturation episode on postoperative day (POD) 3 that required manual bag valve support at 100% fraction of inspired oxygen (Fio₂) and increased sedation. Despite these interventions, he remained severely hypoxic and was subsequently pharmacologically paralyzed. His P/F ratio (the arterial Po₂ from the arterial blood gas divided by the Fio₂) was 70 mm Hg after paralysis, and his ventilator settings were adjusted to volume control ventilation at a tidal volume of 400 mL, rate of 28, positive end-expiratory pressure (PEEP) of 18 cm H₂O, and 100% Fio₂ to provide lung-protective ventilation and optimize oxygenation. When his P/F ratio dropped again, inhaled nitric oxide (iNO) was initiated at 20 parts per million (ppm), with improvement of his P/F ratio to 80–90 mm Hg afterward. His chest x-ray (CXR) that evening was unremarkable. He remained hemodynamically stable despite his significant oxygen requirement.

At this time, the patient demonstrated a significant gap between his end-tidal co₂ (~30 mm Hg) and Paco₂ (~90 mm Hg). With his recent surgery, immobility, and Etco₂-Paco₂ gap, concern was raised for PE, and he was started on empiric therapeutic heparin. He was not transported overnight to the CT scanner to confirm the diagnosis out of concern for patient stability; he was on maximum ventilatory support, and transporting him across the hospital to a CT scanner was felt to be too risky. Furthermore, if he became more hypoxic either en route or at the scanner, limited interventions would be available. Lower extremity duplex scans were performed at the bedside and did not reveal any deep vein thrombosis.

On POD 4, EIT was applied to optimize the patient’s ventilator settings, and it was determined that his lung mechanics would be optimized at a PEEP of 14 cm H₂O (Figure 1). By injecting a hypertonic saline bolus (10 mL of 7% NaCl) via a central venous catheter, EIT also elucidated a large ventilation-perfusion (V/Q) mismatch in the right lung, causing concern for PE (Figure 2). A formal transthoracic echocardiogram (TTE) on POD 4 demonstrated a mildly dilated right ventricle. Pulmonary hypertension was noted, with a right ventricular systolic pressure of 68 mm Hg.

The patient was continued on therapeutic anticoagulation and slowly weaned from his iNO, Fio₂, and PEEP over the next few days, with stable partial pressure of oxygen (Pao₂) levels. On POD 8, he was transported for CT angiography (CTA), which showed multiple filling defects in the right lobe of the lung in both segmental and subsegmental pulmonary arteries. There was an additional filling defect noted in the pulmonary artery for the left upper lobe and lingula. The patient was ultimately extubated on POD 13, downgraded from the ICU on POD 15, and discharged from the hospital on POD 17.

DISCUSSION

EIT is a noninvasive, radiation-free clinical imaging tool to monitor, in real time and at the bedside, the distribution of ventilation.³ It uses an electrode belt that is wrapped around the patient longitudinally between the fifth and sixth intercostal spaces.⁴ The electrode belt measures the changes in resistance that occur across the lung during inhalation and exhalation. As the lungs inflate, the alveolar septa lengthen, which then reduce the transmission of electrical current. This resistance can be measured and correlated with the amount of air that enters the lungs through ventilation. Concurrently, a hypertonic saline bolus can be used to track the distribution of pulmonary blood flow. When taken together, this can map lung ventilation and perfusion defects.⁵ In effect, it is a noninvasive, nonionizing, real-time V/Q scan that the clinician can use at the bedside.

In some animal models, EIT has been used to set ventilator parameters and help optimize gas exchange and respiratory mechanics.⁵ Recent studies have focused on expanding this to humans. One group detailed an algorithm of estimating alveolar collapse and hyperdistention based on EIT using decremental PEEP titration. Their protocol had similar results to data produced by a CT-based protocol for patients with acute respiratory distress syndrome (ARDS).⁶ A later study demonstrated that EIT can optimize PEEP to maximize lung recruitment and minimize ventilator-induced lung injury by avoiding alveoli hyperdistention.⁷ EIT provides near- continuous second-by-second data so patients can be quickly down-titrated from high levels of PEEP without causing adverse outcomes while mapping the ideal PEEP setting.

Given the data to suggest that EIT can be used to detect areas of alveolar collapse, several studies have used EIT as a measure to show improved lung mechanics after proning.^8,9 One case report used EIT to determine the need for proning in patients with ARDS from coronavirus disease 2019 (COVID-19). They measured distribution of ventilation in different regions of the lung before, during, and after proning, and showed that by analyzing the ratios of ventilation at different regions of interest, they could estimate lung compliance to determine which patients may benefit from proning.

A more recent development with EIT has been the use of a hypertonic saline bolus to assess perfusion of pulmonary blood flow, as described by He et al.^10,11 This protocol utilizes injection of hypertonic saline into the pulmonary circulation via a central line during end-expiration. The change in impedance caused by the saline bolus reflects pulmonary arterial flow and is visualized.⁴

A recently published prospective observational study demonstrated that using EIT to measure dead space percentage can accurately detect PE. This outperformed D-dimer with an area under the curve (AUC) measurement of 0.986 compared to D-dimer AUC of 0.502. The authors found that a cutoff value of 30.37% for dead space resulted in 90.9% sensitivity and 98.6% specificity.¹⁰ While this was a small, single-center study to show proof of concept, it did show the potential role of EIT in diagnosing PE at the bedside beyond what had been previously reported in animal models.

When comparing the ability of EIT to diagnose PE to the gold standard of CTA, there are important limitations. Other disease processes create dead space, which can reduce the specificity of this diagnostic modality. It is important to note that no randomized controlled trials have shown an advantage of EIT over existing imaging studies. However, an advantage of EIT is that it can be done portably at the bedside. More research is needed on the effect of patient body mass index (BMI) on interfering with EIT measurements. A small 22-person study found differences in anthropometric variability to account for a 1.3% difference, at most.¹²

Our case showed that EIT has the potential to be used to diagnose PE in the postoperative patient population. Starting anticoagulation in a postoperative patient can be a risky endeavor. However, for many of these patients in the ICU, transport to a CT scanner may impose just as much risk. The data from EIT helped bolster the decision to treat for a presumptive PE in our patient. Another report advocated for the use of EIT via a similar methodology to diagnose PE in a patient with COVID-19.¹³ However, further work is needed to determine its general applicability as well as to implement standardized protocols and cutoff values for its use as a diagnostic tool.

EIT has clinical utility in critical care setting for diagnostic purposes as well as ventilatory optimization. Its use needs controlled trial validation, but as clinicians become more familiar with its use and limitations, it may become a more utilized and helpful tool in the management of critically ill patients.

REFERENCES