Use of thrombolytic therapy beyond current recommendations for acute ischaemic stroke

In Chapter 1, I introduce ischaemic stroke, thrombolytic therapy, thrombolysis trials and then discuss the rationale for exclusion criteria in stroke thrombolysis guidelines.In Chapter 2, I describe methods for examining outcomes in patients that are currently recommended for exclusions from receiving alteplase for acute ischaemic stroke. In Chapter 3, I examine Virtual International Stroke Trials Archive (VISTA) data to test whether current European recommendation suggesting exclusion of elderly patients (older than 80 years) from thrombolysis for acute ischaemic stroke is justified. Employing non-randomised controlled comparison of outcomes, I show better outcomes amongst all patients (P < 0.0001; OR, 1.39; 95% CI, 1.26 to 1.54), young patients (P < 0.0001; OR, 1.42; 95% CI, 1.26 to 1.59) and the elderly patients (P = 0.002; OR, 1.34; 95% CI, 1.05 to 1.70). Odds Ratios are consistent across all age deciles > 30 years. Outcomes assessed by National Institutes of Health Scale (NIHSS) score and dichotomised modified Rankin Scale score are consistently similar. In Chapter 4, I compare thrombolysed patients in Safe Implementation of Thrombolysis in Stroke International Stroke Thrombolysis Register (SITS-ISTR) with VISTA non-thrombolysed patients ("comparators" or "controls") and test exactly similar question as in Chapter 3. Distribution of scores on modified Rankin scale are better amongst all thrombolysis patients than controls (odds ratio 1.6, 95% confidence interval 1.5 to 1.7; Cochran-Mantel-Haenszel P<0.001). Association occurs independently amongst patients aged ≤80 (0R 1.6, 95%CI1.5 to 1.7; P<0.001; n=25,789) and in those aged >80 (OR 1.4, 95% CI 1.3 to 1.6; P<0.001; n=3439). Odds ratios are consistent across all 10 year age ranges above 30, and benefit is significant from age 41 to 90; dichotomised outcomes (score on modified Rankin scale 0-1 v 2-6; 0-2 v 3-6; and 6 (death) versus rest) are consistent with the results of ordinal analysis. These findings are consistent with results from VISTA reported in Chapter 3. Age alone should not be a criterion for excluding patients from receiving thrombolytic therapy.In Chapter 5, I employ VISTA data to examine whether patients having diabetes and previous stroke have improved outcomes from use of alteplase in acute ischaemic stroke. Employing a non-randomised controlled comparison, I show that the functional outcomes are better for thrombolysed patients versus nonthrombolysed comparators amongst non-diabetic (P < 0.0001; OR 1.4 [95% CI 1.3-1.6]) and diabetic (P = 0.1; OR 1.3 [95% CI1.05-1.6]) patients. Similarly, outcomes are better for thrombolysed versus nonthrombolysed patients who have not had a prior stroke (P < 0.0001; OR 1.4 [95% CI1.2-1.6]) and those who have (P = 0.02; OR 1.3 [95% CI1.04-1.6]). There is no interaction of diabetes and prior stroke with treatment (P = 0.8). Neurological outcomes (NIHSS) are consistent with functional outcomes (mRS). In Chapter 6, I undertake a non-randomised controlled comparison of SITS-ISTR data with VISTA controls and examine whether patients having diabetes and previous stroke have improved outcomes from use of alteplase in acute ischaemic stroke. I show that adjusted mRS outcomes are better for thrombolysed versus non-thrombolysed comparators amongst patients with diabetes mellitus (OR 1.45[95% CI1.30-1.62], N=5354), previous stroke (OR 1.55[95% CI1.40-1.72], N=4986), or concomitant diabetes mellitus and previous stroke (OR 1.23 [95% CI 0.996-1.52], P=0.05, N=1136), all CMH p<0.0001. These are comparable to outcomes between thrombolysed and non-thrombolysed comparators amongst patients suffering neither diabetes mellitus nor previous stroke: OR=1.53(95%CI 1.42-1.63), p<0.0001, N=19339. There are no interaction between diabetes mellitus and previous stroke with alteplase treatment (t-PA*DM*PS, p=0.5). Present data supports results obtained from the analyses of VISTA data in chapter 5. There is no statistical evidence to recommend exclusion of patients with diabetes and previous stroke from receiving alteplase.In Chapter 7, I examine VISTA data to test whether exclusion of patients having a mild or severe stroke at baseline would be justified. Stratifying baseline stroke severity for quintiles of NIHSS scores, I observe that there are significant associations of use of alteplase with improved outcomes for baseline NIHSS levels from 5 to 24 (p<0.05). This association lose significance for baseline NIHSS categories 1 to 4 (P = 0.8; OR, 1.1; 95% CI, 0.3-4.4; N = 8/161) or ≥ 25 (P = 0.08; OR, 1.1; 95% CI, 0.7-1.9; N = 64/179) when sample sizes are small and confidence interval wide. These findings fail to provide robust evidence to support the use of alteplase in the mild or severe stroke patients, though potential for benefit appears likely.In Chapter 8, I present a meta-analysis of trials that investigated mismatch criteria for patients’ selection to examine whether present evidence supports delayed thrombolysis amongst patients selected according to mismatch criteria. I collate outcome data for patients who were enrolled after 3 hours of stroke onset in thrombolysis trials and had mismatch on pre-treatment imaging. I compare favourable outcome, reperfusion and/or recanalisation, mortality, and symptomatic intracerebral haemorrhage between the thrombolysed and non-thrombolysed groups of patients and the probability of a favourable outcome among patients with successful reperfusion and clinical findings for 3 to 6 versus 6 to 9 hours from post stroke onset. I identify articles describing the DIAS, DIAS II, DEDAS, DEFUSE, and EPITHET trials, giving a total of 502 mismatch patients thrombolysed beyond 3 hours. The combined adjusted odds ratios (a-ORs) for favourable outcomes are greater for patients who had successful reperfusion (a-OR=5.2; 95% CI, 3 to 9; I2=0%). Favourable clinical outcomes are not significantly improved by thrombolysis (a-OR=1.3; 95% CI, 0.8 to 2.0; I2=20.9%). Odds for reperfusion/recanalisation are increased amongst patients who received thrombolytic therapy (a-OR=3.0; 95% CI, 1.6 to 5.8; I2=25.7%). The combined data show a significant increase in mortality after thrombolysis (a-OR=2.4; 95% CI, 1.2 to 4.9; I2=0%), but this is not confirmed when I exclude data from desmoteplase doses that are abandoned in clinical development (a-OR=1.6; 95% CI, 0.7 to 3.7; I2=0%). Symptomatic intracerebral haemorrhage is significantly increased after thrombolysis (a-OR=6.5; 95% CI, 1.2 to 35.4; I2=0%) but not significant after exclusion of abandoned doses of desmoteplase (a-OR=5.4; 95% CI, 0.9 to 31.8; I2=0%). Delayed thrombolysis amongst patients selected according to mismatch imaging is associated with increased reperfusion/recanalisation. Recanalisation/reperfusion is associated with improved outcomes. However, delayed thrombolysis in mismatch patients was not confirmed to improve clinical outcome, although a useful clinical benefit remains possible. Thrombolysis carries a significant risk of symptomatic intracerebral haemorrhage and possibly increased mortality. Criteria to diagnose mismatch are still evolving. Validation of the mismatch selection paradigm is required with a phase III trial. Pending these results, delayed treatment, even according to mismatch selection, cannot be recommended as part of routine care.In Chapter 9, I summarise the findings of my research, discuss its impact on the research community, and discuss weaknesses inherent in registry data and limitation of statistical methods. Then, I elaborate the future directions I may take to further research on the theme of this thesis.