The use of separate‐level neuraxial anaesthesia for caesarean delivery in a patient with a history of spinal tuberculosis

Abstract

We present the case of a 33‐year‐old parturient who required caesarean delivery at 31 weeks' gestation. She had a history of degenerative disease of the lumbar spine secondary to tuberculosis, acquired as a child in India. Her complex medical history also included ischaemic heart disease and obstructive sleep apnoea, and due to this general anaesthesia was considered to be of high risk. However, regional anaesthesia also posed significant challenges because magnetic resonance imaging of the spine showed a partial collapse with subsequent fusion of second and third lumbar vertebral bodies with thoracolumbar kyphosis. Neuraxial anaesthesia was performed with ultrasound guidance for determining levels and depth of epidural space. An epidural was inserted at the T12–L1 interspace and a spinal anaesthetic block was placed at L4–L5. Delivery and recovery were uneventful. This case highlights the safe and effective use of neuraxial anaesthesia in an asymptomatic patient with treated spinal tuberculosis as well as the usefulness of high‐quality imaging of the spine in the decision to perform neuraxial anaesthesia.

Introduction

Spinal tuberculosis (TB) is an infection of the vertebrae, also known as Pott's disease. It is evident throughout human history having been documented in spinal remains from the Iron Age in Europe and in ancient mummies from Egypt. Pott's disease results from haematogenous spread of TB from distant sites, and is often pulmonary in origin. The infection then spreads from two adjacent vertebrae into the adjoining intervertebral disc space. If only one vertebra is affected, the disc is preserved, but if two are involved, the disc, which is avascular, collapses. This, combined with eventual vertebral collapse, leads to kyphoscoliosis and potential spinal cord damage. Neuraxial anaesthesia in a patient with a prior history of spinal TB can therefore present a technical challenge to the anaesthetist as it can distort the anatomy, prevent the successful location of the epidural or subarachnoid space and result in an unpredictable distribution of the injectate. Radiological assessment in the form of magnetic resonance imaging (MRI) and ultrasound can be useful in the identification of suitable spaces for needle placement. Though TB is uncommon in the UK, it is highly prevalent worldwide but despite its prevalence, the literature on the safety of neuraxial anaesthesia in patients with a history of treated spinal TB is sparse. In this report, we add to that literature by describing the anaesthetic management for elective caesarean section in a patient with a history of Pott's disease.

Report

A 33‐year‐old gravida 2 para 1 patient required the pre‐term caesarean delivery of a small‐for‐gestational‐age fetus. Her obstetric history was of one previous caesarean delivery under spinal anaesthetic 7 years previously in another country. Her medical history was complex and included: body mass index of 40 kg.m⁻²; essential hypertension; type 2 diabetes; obstructive sleep apnoea requiring nocturnal continuous positive airway pressure and coronary artery disease. She was admitted at 28 weeks' gestation for in‐patient fetal monitoring due to a deterioration in fetal growth as seen on ultrasound, and during this admission, an anaesthetic assessment was carried out. Her medications at the time of this assessment included: aspirin 75 mg; clopidogrel 75 mg; enoxaparin 60 mg once daily; bisoprolol; and methyldopa. Her diabetic control required large doses of insulin and metformin.

Fifteen months previously, she had suffered a non‐ST elevation myocardial infarction which was treated with drug eluting coronary artery stenting and dual anti‐platelet therapy. Four months after this, she developed chest tightness and had a coronary angiogram which showed in‐stent thrombosis and she did not receive any further stenting. A transthoracic echocardiograph performed at 17 weeks' gestation demonstrated good biventricular systolic function with no obvious regional wall motion abnormality or significant valvular abnormalities. She also had a history of spinal TB as a child in India. Her memory of this was limited but she recalled receiving treatment for it and denied any neurological deficit. On examination, she had midline scar tissue overlying the second, third and fourth lumbar vertebrae. An MRI spine was performed during her inpatient stay and revealed a partial collapse with subsequent fusion of second and third lumbar vertebral bodies with thoracolumbar kyphosis (Fig. 1) as well as a wedge collapse of the seventh cervical vertebral body. The spinal cord was seen to kink around the kyphotic curves but there was no evidence of cord compression. In anticipation of a likely urgent premature delivery due to a tail off in growth velocity on ultrasound over a number of weeks, it was decided to omit her clopidogrel from 29 weeks' gestation.

Figure 1.

Magnetic resonance imaging scan of the spine. The red arrow indicates the fusion of second and third lumbar vertebral bodies, with resultant thoracolumbar kyphosis.

The decision to proceed with elective caesarean delivery was made at 31 weeks and 2 days, following omission of her aspirin and enoxaparin for 24 h. Separate level combined spinal and epidural anaesthesia was performed with pre‐procedural ultrasound scanning used to determine the optimum spaces and the depth of epidural space, and verify the spinal levels. With the patient in the sitting position, an epidural was sited at the T12–L1 interspace using a standard 16G Tuohy needle, and an intrathecal injection was performed at the L4–L5 interspace using a 25G Whitacre needle. The spinal dose was 1.8 ml hyperbaric bupivacaine 0.5% with fentanyl 25 μg. The block was tested with ethyl chloride spray as is standard in our institution. A slow incremental dose of levobupivacaine 0.5% was given via the epidural catheter to a total of 5 ml in order to achieve a block for surgery. Haemodynamic stability was maintained with a phenylephrine infusion (100 μg.ml‐⁻¹), commenced following the intrathecal injection and titrated to effect. A further 2 ml epidural bolus of levobupivacaine 0.5% and 2.5 mg diamorphine was administered following delivery of the baby. The patient was transferred to the high dependency unit and had an uneventful postoperative course. She was recommenced on her treatment dose of enoxaparin after the caesarean section and following discussion with the obstetric and medical teams, her antiplatelet medication was recommenced after 72 h.

Discussion

Though TB is uncommon in the UK, it is highly prevalent worldwide and despite its prevalence, the literature on the safety of neuraxial anaesthesia in patients with a history of treated spinal TB is sparse. As far as we are aware, there are reports on the management of five obstetric patients with a history of spinal TB and kyphoscoliosis available in the literature [1, 2, 3, 4, 5]. Four of these come from the anaesthetic literature [1, 2, 3], and one from the obstetric literature with some mention of anaesthesia management [4]. In all but one of the reported cases, the patient arrived at the care of the anaesthetist in an emergency setting. The second patient reported by Dabrowska et al. was seen antenatally and an MRI was arranged [2]. In one case, a patient underwent general anaesthesia for caesarean delivery, the decision to do so being based on the complexity of her kyphoscoliosis and the lack of previous evaluation of her spine [4]. All other patients received neuraxial anaesthesia with no reported complications afterwards, though none of the cases described a technique of combining spinal anaesthesia with an epidural. Though our patient was broadly similar to these patients in that, she had a spinal deformity related to TB, she was complicated by the fact that she required anticoagulation and had numerous medical co‐morbidities.

The use of combined spinal and epidural anaesthesia in the obstetric setting was first described by Brownridge in 1981 who described epidural catheter placement at the L1–L2 interspace followed by subarachnoid block at L3–L4 [5]. The use of separate interspaces for combined spinal and epidural is less popular today than the needle‐through‐needle technique which we routinely use in our institution. The theoretical risk of using separate interspaces is that insertion of the spinal needle may damage the epidural catheter. This risk was minimised in our patient by performing a low thoracic epidural, therefore separating the blocks by anatomical distance. Our decision to perform separate level neuraxial anaesthesia was primarily based on upon concern that the local anaesthetic may not spread beyond the mid‐lumbar kyphosis in our patient as well as consideration of her underlying cardiac issues and the ability to titrate the dose of local anaesthetic to effect. The anatomical concern was based on reviewing the MRI images as well as the very obvious scar tissue mass over the area.

Magnetic resonance imaging is the cross‐sectional imaging modality of choice in pregnancy as it avoids the doses of ionising radiation associated with computerised tomography. It is also the benchmark modality for evaluating both spinal degenerative disorders and spinal infections [6]. In our patient, it allowed us to identify the level of the lesion and assess the spinal cord and was therefore useful in planning the type and location of the neuraxial blocks. Ultrasound scanning of her spine also played a role in the decision of where to place the epidural; our ultrasound assessment demonstrated grossly abnormal interspaces at L1–L2 and L3–L4, consistent with MRI findings. It also showed large variations in the depth to the epidural space at different levels. At L4–L5, the depth to the epidural space was 8 cm and at T12–L1 it was 4 cm. Based on these findings we decided to place the epidural at T12–L1. The combination of MRI and ultrasound imaging were useful in the management of this patient. This is consistent with Vercauteren et al.'s review of neuraxial techniques in obstetric patients with pre‐existing spinal pathology which concludes that ultrasound technology may help the anaesthetist to determine the midline and the best lumbar interspace and approach [7].

Our patient was admitted as an in‐patient for fetal monitoring before her delivery. We were therefore fortunate that this allowed time to assess, investigate and formulate a plan with our obstetric and medical colleagues in consultation with a neuroradiologist. Though delivery in this case was by elective caesarean section, a plan was also in place should there have been a deterioration in the fetal condition and more urgent delivery was required; this involved performing a general anaesthetic, due to the fact that she would have been anticoagulated. Based on her medical history, she would have required invasive arterial monitoring and a modified rapid sequence induction in order to maintain cardiovascular stability.

This case highlights the importance of imaging in the planning of complex regional anaesthesia, and of referral and liaison between specialties. By having the opportunity to plan the procedure carefully, we were able to provide high‐quality anaesthetic care to a patient for whom the risks and benefits of general versus neuraxial anaesthesia needed to carefully be considered, optimise the timing of her caesarean delivery and safely manage the dosing of her anticoagulant medications.