Editor, Friedreich's ataxia is the most frequently inherited ataxia (1:29 000 cases in a white population). When patients affected by this disorder are scheduled for a surgical procedure, important anaesthetic issues have to be taken into account to optimise perioperative management. We report the case of a 35-year-old woman who was scheduled for elective correction of right cavus foot. Written consent was obtained from the patient to publish her clinical case. Other than the disorders associated with Friedreich's ataxia diagnosed 13 years earlier, she had no significant medical history. Clinical examination revealed a thoracic scoliosis, pes cavus, ocular ataxia, hypoaesthesia with decreased sensation to vibration more obvious in the lower limbs and cerebellar syndrome (gait ataxia, dysmetria, dysdiadochokinesis, dysarthria and intention tremor). The EMG and somatory-evoked potentials did not show major alterations other than mild signs of denervation. No anomaly was visible on MRI. Genetic studies had identified two different known haplotypes of the Frataxin (FXN) gene associated with Friedreich's ataxia. At preoperative assessment, the results of routine blood tests were normal. Right bundle branch block was present on the ECG, but an echocardiogram and cardiac MRI were normal. Pulmonary function tests showed a mild restrictive pattern with no blood gas disturbance or chest radiograph anomalies other than the scoliosis. The intervention (triple arthrodesis with metatarsal osteotomies and Kirschner wire insertions) was performed with a tourniquet under combined spinal-epidural (CSE) anaesthesia. With the patient in a sitting position, the epidural space was located using an 18G Tuohy needle at L4 to L5 with a loss of resistance to air technique. Intrathecal hyperbaric bupivacaine 13 mg and fentanyl 10 μg were administered through a 27G Whittacre spinal needle inserted through the Tuohy needle. A 21G multiperforated epidural catheter was then inserted (Perifix set; B-Braun). A sensory level to T8 was reached with no haemodynamic or respiratory disturbances. The intraoperative period, which lasted 90 min, was uneventful, and the patient was haemodynamically stable during the procedure. At the end of surgery, the patient was transferred to the postanaesthesia care unit. After 1 h, the motor block had recovered and the patient was discharged to the ward with a continuous epidural infusion of levobupivacaine 0.125% at 5 ml h−1 for the treatment of postoperative pain, which was discontinued on day 3. The patient was discharged from hospital on day 3 with no respiratory or neurological complications. Friedreich's ataxia is caused by hyperexpansion of trinucleotide GAA in the FXN gene located in chromosome 9, which provokes a deficiency of frataxin, a mitochondrial protein responsible for iron metabolism and oxidative status. This deficiency leads to iron overload and consequent mitochondrial dysfunction,1 which explains the neurological presentations of the disease, cardiomyopathy, alterations of glucose metabolism with hyperglycaemia and the scoliosis typically described in the disease. These patients are considered at high risk for anaesthesia, despite their young age (disease onset usually occurs before 25 years of age, with a median age of death of 35 years and a better prognosis for women). Concerns about the possible risk of worsening of neurological symptoms using central neuraxial blockade were discussed, but the safe use of spinal and epidural anaesthesia has been described in patients with Friedreich's ataxia, usually for labour analgesia and caesarean section.2 In the present case, a CSE technique was chosen to allow the administration of top-up doses and to avoid general anaesthesia in case of a prolonged operation, and to provide postoperative analgesia. The duration of surgery was expected to be more than 90 min, but the patient did not present with cardiomyopathy or severe neuromuscular dysfunction, and a relatively large dose of bupivacaine was used to avoid the need for an intraoperative epidural top-up dose. The CSE technique can also be used to administer low doses of intrathecal local anaesthetic, with a lower incidence of sympathetic blockade, compensating for the shorter duration of the blockade by the administration of titrated top-up epidural doses if required. Despite the unpredictable risk of autonomic dysfunction in this case, we adopted a strategy with a higher dose of local anaesthetic because she did not present with any previous severe haemodynamic alterations.3 She required no analgesia other than the epidural infusion in the postoperative period and was haemodynamically stable throughout her hospital stay. General anaesthesia can jeopardise the outcome of patients with Friedreich's ataxia due to intraoperative and postoperative respiratory disorders caused by thoracic kyphoscoliosis, associated with restrictive respiratory function. In addition, the use of nondepolarising muscle relaxants has been reported to have an unpredictable response, with cases of prolonged block leading to difficult weaning from the ventilator, longer recovery and a prolonged hospital stay.4 Reports of hyperkalaemia following administration of succinylcholine in patients with denervated muscle suggest that it would be preferable to avoid its use. If muscle relaxation is essential, a nondepolarising muscle relaxant can be used with care, with monitoring of the neuromuscular block.5 The mitochondrial dysfunction suggests that propofol-based total intravenous anaesthesia (TIVA) should not be used in these patients because of its depressant effects on mitochondrial metabolism, and its possible implication in propofol-infusion syndrome. These concerns make the use of regional anaesthesia preferable if general anaesthesia can be avoided.6 Glucose-containing solution with electrolytes should be used to ensure that the mitochondria receive enough glucose to fuel the respiratory chain without relying on fatty acid oxidation; the blood glucose concentration should thus be monitored to avoid hyperglycaemia due to the metabolic disturbance and impaired glucose metabolism associated with Friedreich's syndrome. In conclusion, although CSE anaesthesia resulted in a satisfactory outcome, avoiding the complications of general anaesthesia, the use of a peripheral block would have been a very good alternative option in this case.7 Ultrasound-guided injection of local anaesthetic with insertion of a perineural catheter would have avoided the risks of a central block, but muscle degeneration can modify echogenicity and make the identification of neural structures more difficult, as has been found in Duchenne's muscular dystrophy. Acknowledgements relating to this article Assistance with the study: none. Financial support and sponsorship: none. Conflict of interest: none.