The implications of the GALA trial General Anaesthesia

The implications of the GALA trial: General Anaesthesia vs Local (regional) Anaesthesia for Carotid Surgery 3 rd UK Stroke Forum Conference, Harrogate, December 2008 Michael Gough, Leeds and Charles Warlow, Edinburgh for the GALA collaborators Funding: Health Foundation, ESVS

The sort of calculation that one can do in one’s head… • For >70% symptomatic stenosis Risk of surgery: 5% stroke/death within 30 days Risk of ipsilateral ischaemic stroke without surgery: 20% at two years Risk of death/another sort of stroke within two years: very low Risk of ipsilateral ischaemic stroke after successful surgery: “zero” • Calculation Absolute risk reduction in stroke from surgery: 15% (20 - 5) Number-needed-to-operate to prevent a stroke = 6 (100/15) Therefore 1 in 6 patients benefit from surgery, 5 do not Funding: Health Foundation, ESVS

Interpretation • If number-needed-to-operate = 6 patients, to make surgery a ‘better buy’ (reduce number-needed-to-operate): Identify patients with higher ipsilateral stroke risk without operation Safer investigation (angiography) Safer surgery (identify low surgical risk) Safer anaesthesia: GALA Funding: Health Foundation, ESVS

General (GA) or Local Anaesthesia (LA) for carotid surgery: pros and cons • Advantages to LA ‘Awake neurological testing’ during carotid clamping = ↓shunting Preserves autoregulation • Potential benefits of LA ? ‘safer’ in high risk elderly ‘vascular’ patients ? less ‘stress’ response to surgery ? better postoperative pain relief ? earlier mobilisation, less traumatic = QOL, less expensive v GA • Possible disadvantages of LA More traumatic for the patient and the surgeon Hurried surgery Conversions (LA to GA) can be problematic Patient might prefer GA Funding: Health Foundation, ESVS

Cochrane Review of LA v GA for carotid surgery: non-randomised, stroke and death Rerkasem K, Bond R, Rothwell PM. Cochrane Database of Systematic Reviews 2004; 2: CD 000126 Funding: Health Foundation, ESVS

Cochrane Review of LA v GA for carotid surgery: randomised, stroke and death Rerkasem K, Bond R, Rothwell PM. Cochrane Database of Systematic Reviews 2004; 2: CD 000126 Funding: Health Foundation, ESVS

Rationale for GALA • Good theoretical reasons to prefer LA over GA for CEA but …. … “beautiful hypotheses can be destroyed by ugly facts” (Thomas Huxley) • Cochrane Review encouraging but… non-randomised studies likely to be biased randomised trials too small ‘stroke and death’ are not the only outcomes of interest • Variation in practice of carotid surgery over time • No good evidence for LA vs GA in other forms of surgery Funding: Health Foundation, ESVS

What happened next? 1997: CPW, MJG Steering Committee Protocol MREC Trial Co-ordinator Funding 1999: Pilot 20 UK Centres 2003: Main Trial Funding: Health Foundation, ESVS

Design of GALA • Randomised, partially blinded two arm trial, intention-to–treat analysis • Uncertainty principle • Pragmatic non-restrictive protocols (except shunt in LA) • Management Leeds: surgical and anaesthetic leadership Edinburgh: trial Management York: health economics • Target: 5000 patients • Follow up at: Ø hospital discharge, 7 days post operative, or death Ø one month: ‘blind’ stroke physician/neurologist (phone if necessary) Ø one month: QOL questionnaire (UK only) Ø one year: questionnaire to patients re stroke/MI Funding: Health Foundation, ESVS

Why 5000 patients? • Assume 7. 5% incidence of primary outcome at 30 days • Achieve one third reduction in risk to 5% (> 90% power at 5%) • Analysis intention-to-treat • Primary outcome: Stroke (including retinal infarct), myocardial infarction (MI), death • Secondary outcomes: Alive and stroke/MI free at one year QOL at 30 days (UK only) Surgical complications (haematoma, re-opn, cranial nerve palsy etc) Length of stay (intensive care, high dependency, total) Cost Funding: Health Foundation, ESVS

Eligibility for the GALA Trial • Experienced surgeons (>15 carotid endarterectomies per annum) • Local ethics committee approval • Any patient requiring carotid surgery (symptomatic or asymptomatic stenosis) • Usual management, except shunts during LA only if indicated by awake testing • Uncertainty • No patient preference Funding: Health Foundation, ESVS

3526 patients from 95 GALA centres in 24 countries CHINA AUSTRALIA Funding: Health Foundation, ESVS

3526 randomised (95 centres, 24 countries) GA 1753 allocated: 1628 GA 31 no anaesthesia - 92 cross-over 2 unknown LA 1773 allocated: 1655 LA 41 no anaesthesia - 75 cross-over 2 unknown 99. 9% FU 1752 for primary outcome (No FU = 1, Incomplete = 20) 1771 for primary outcome (No FU = 2, Incomplete = 19) Funding: Health Foundation, ESVS

Baseline data General Local Age 70 (sd 9) 69 (sd 9) Male 1232 (70%) 1256 (71%) Asymptomatic stenosis 685 (39%) 677 (38%) Mean % stenosis 81 (sd 11) Contralateral ICA occlusion 150 (9%) 160 (9%) Smoking, peripheral arterial disease, coronary artery disease, atrial fibrillation, diabetes, blood pressure all equal Funding: Health Foundation, ESVS

Compliance General Local No anaesthesia Stroke or death before operation 2 2 Carotid artery occlusion 8 8 Too ill (not carotid), Stenosis too mild, stent 5 12 Patient refused 9 31 13 41 Conversion post- anaesthesia, pre-op Patient’s decision 6 Problem with position on table etc 3 Patient deteriorated after local block 8 Conversion after start of surgery Pain, discomfort, anxiety, claustrophobia 34 Physiological instability, protracted surgery 11 Neurological deterioration on cross-clamping 7 Funding: Health Foundation, ESVS

Compliance – cross-overs General (n=92) Local (n=75) Medical decision 41 20 Administrative issues 15 9 Patient’s decision 29 44 Reason unknown 7 2 Reasons: Funding: Health Foundation, ESVS

Primary outcome events Intention-to-treat 5% 4% 10 4 5 9 3% 2% 70 66 General 84/1752 (4. 8%) MI (fatal or non-fatal) Stroke (fatal or non-fatal) Local 80/1771 (4. 5%) 1% 0% Other deaths Funding: Health Foundation, ESVS

Primary outcome events Intention to treat Stroke 3 (-10 to +16) MI -4(-8 to +2) Death (any cause) 4 (-3 to +12) Stroke or death 4 (-9 to +18) Stroke, MI or death 3 (-11 to +17) -20 -10 0 10 20 Events prevented/1000 (95% CI) Favours General Favours Local Funding: Health Foundation, ESVS

Strokes within 30 days of CEA . infarct haemorrhage unknown 70 Number of patients 80 60 50 40 30 20 10 0 Preop 0 1 2 3 4 5 -7 8 -14 15 -21 22 -30 Days since endarterectomy Funding: Health Foundation, ESVS

Subgroup analysis on primary outcome Contralateral carotid occlusion Favours LA Favours GA Funding: Health Foundation, ESVS

Secondary outcomes No definite differences (GA v LA): Length of stay Duration of surgery Trainee v consultant Asymptomatic v symptomatic UK v others Cranial nerve injury Wound haematoma Chest infection Quality of life at one month Outcome at one year Cost Funding: Health Foundation, ESVS

Survival analysis Free of stroke, MI and death Funding: Health Foundation, ESVS

Limitations of GALA • Lack of power Sample size, outcome events • Lack of complete blinding • Cross-overs pre-op (5%), conversions LA GA (4%) • Lack of standardisation of anaesthetic and surgical protocols BP in the GA group, Patching: 42% LA v 50% GA • The surgical risk model did not work • Took too long, would have failed without the non-UK centres Funding: Health Foundation, ESVS

UK and Non UK Centres Number of patients randomised/year 900 800 Non UK UK 700 Patients 600 500 400 300 200 100 0 1999 2000 2001 2002 2003 2004 2005 2006 2007 Funding: Health Foundation, ESVS

Recruitment in Carotid Surgery Trials 4000 3526 Number of Patients 3500 3024 3000 2500 3120 2267 2000 1500 1000 500 0 NASCET ECST ACST 1 GALA Funding: Health Foundation, ESVS

Limitations of local anaesthesia • Unable to tolerate • Additional sedation and analgesia • Conversion to GA • Stress & anxiety may cardiac events • Injury to surrounding structures • More peri-operative strokes may be due to embolism • Modern GA safer/less stressful Funding: Health Foundation, ESVS

Putting GALA into context Stroke & death OR (95% CI) Meta-analysis of 7 earlier RCTs 0. 62 (0. 24 to 1. 59) GALA 0. 88 (0. 64 to 1. 23) Meta-analysis including GALA 0. 85 (0. 63 to 1. 16) Favours Local Favours General Funding: Health Foundation, ESVS

Putting GALA into context Death OR (95% CI) Meta-analysis of 7 earlier trials 0. 23 (0. 05 - 1. 01) GALA 0. 72 (0. 40 - 1. 30) Meta-analysis including GALA 0. 62 (0. 36 – 1. 07) Favours Local Favours General Funding: Health Foundation, ESVS

Conclusions • Little difference in patient outcomes regardless of GA or LA • Surgical teams should be able to offer both LA & GA • The individual choice should be determined by the patient’s medical need and personal preference • Trials like GALA could and should be done more quickly, but will have to be multinational • Regulations make trials increasingly difficult to do, and more expensive • The cost-effectiveness of carotid endarterectomy would be improved more dramatically by shortening the time from symptoms to surgery Funding: Health Foundation, ESVS

The GALA Trial A collaboration Healthcare Foundation Vascular Surgeons throughout Europe Funding: Health Foundation, ESVS