This case was challenging for several reasons. Her limited range of neck movements posed a high degree of technical difficulty for tracheal intubation. Although this could be circumvented by using a fiberoptic technique or supraglottic airway device, the presence of a retrosternal goiter occupying the anterior mediastinum was a much greater concern, since it could cause total airway obstruction under general anaesthesia. Moreover, pneumonia had reduced her respiratory reserve, thus limiting the duration of intubation apnoea. In the event of total airway obstruction, the default option of rigid bronchoscopy as a rescue plan was contraindicated due to restricted neck movement. Therefore, emergency sternotomy or pre-emptive placement of extracorporeal membrane oxygenator were the only viable options for airway and ventilatory support. Therefore, general anaesthesia was deemed as a very high-risk procedure.

The hip and thigh are innervated by branches of the lumbrosacral plexus, namely, the femoral nerve, obturator nerve, and lateral femoral cutaneous nerve from the lumbar plexus, and the sciatic nerve, posterior cutaneous nerve of the thigh, and superior gluteal nerve from the sacral plexus. Performing regional anesthesia by blocking the plexus or all branches is technically challenging and entails a high risk of local anaesthetic systemic toxicity due to the high volume of local anaesthetic that would be required to allow adequate coverage. Furthermore, in this case, use of sedation as a supplement could lead to a reduction in muscle tone, resulting in total airway obstruction. Therefore, regional anaesthesia was not considered a safe option for the patient. It has been reported that asymptomatic patients with tracheal stenosis of more than 50% of the residual cross area, as in our patient, are considered to be at high risk of developing intraoperative airway problems⁵.

Pre-procedural imaging of the spine in this particular patient group has been described in a few studies^6,7,8 CT and MRI are more sensitive than plain radiographs for assessing the involvement of the spine⁸. Three-dimensional CT is best for demonstrating interspinous and interlaminar spaces, whereas MRI is best for demonstrating soft tissue inflammation⁶. Compared to US, 3D-CT is non-operator-dependent, thus providing an objective means of identifying unfused interlaminar spaces for a preoperative assessment regarding the feasibility of SA. Various SA techniques have also been explored, including US-guided⁸ and fluoroscopy-guided^9,10 Taylor’s approach¹¹ and mini-laminectomy¹². Both US and fluoroscopy are minimally invasive and are familiar techniques in our institute. However, the latter requires the patient to be laid in a prone position, which can be distressing and potentially hazardous in a patient with a fractured hip and an obstructive retrosternal goitre. Moreover, fluoroscopy exposes the patient and operators to radiation, thereby making fluoroscopy-guided SA a less favorable choice.

The potential of US in predicting the ease of performing neuraxial anesthesia has been evaluated in two cohort studies^13,14. These were based on the premise that the ability to visualise the vertebral canal corresponds to the size of the interlaminar space, thus reflecting the ease with which it may be penetrated. However, the angle of the US scan in attaining the most optimal view and, therefore, the path of needle advancement may not be accurately replicated during the actual procedure. This may subsequently lead to failure, especially in patients with challenging anatomies. Real-time US-guided SA overcomes this limitation by allowing manipulation of the needle under direct vision. This advantage is illustrated in our case by the fact that redirection of the needle was not required, and SA was conducted in a single attempt once the optimal view was acquired.

Although US has been shown to allow identification of interlaminar spaces, epidural/intrathecal spaces, and adjacent structures¹⁵, its accuracy is affected by multiple factors, including the quality of US equipment, angle of US beam penetration, patient positioning, and operator technique. These factors may account for the discrepancy in measured sizes of the interlaminar spaces by US and 3D-CT. Another caveat of this technique is the technical difficulty in maintaining the desired view on US while manipulating the needle. We opted to have two operators to perform each of these tasks; however, several single-operator techniques have been described in other studies. Visualising the needle on US can be challenging because of the steep angle of needle advancement and the low-frequency transducer that is typically used in US. Ideally, echogenic spinal needles should be used; however, since they were not available in our institute, the 18G Tuohy needle was chosen for better visibility rather than non-echogenic spinal needles.

In conclusion 3D-CT provides an objective way of identifying unfused interlaminar spaces to assess the feasibility of a SA while real-time US-guided SA offers the advantage of allowing manipulation of needle under direct vision without exposure to radiation.