Statistics in Practice Part II - Tagged.pdf

Professor Abdel Douiri Statistics in Practice School of Life Course & Population Sciences Aims for today (learning objectives) ‘To understand the role of statistics in real research through statisticians’ own collaborative research’ (Not to learn statistics as some metho...

Professor Abdel Douiri Statistics in Practice School of Life Course & Population Sciences Aims for today (learning objectives) ‘To understand the role of statistics in real research through statisticians’ own collaborative research’ (Not to learn statistics as some methods are very advanced but to see how it is used. Not for examining) Post-Stroke Cognitive Impairment 2 Collaborative research: Stroke epidemiology Douiri et al. Stroke 2013;44:138-45 Prevalence of post stroke cognitive impairment: South London Stroke Register 1995-2010 Douiri et al. Circulation 2013; 128:1341-1348 Long-term effects of secondary prevention on cognitive function in stroke patients 3  The team Dr Abdel Douiri Professor in Medical Statistics and Clinical Trials Prof Tony Rudd, CBE Professor of Stroke Medicine Prof Charles Wolfe Professor of Public Health 4 Outline Background Objectives and study design Results & clinical perspective Conclusion & future work 5 Background 6 Stroke Stroke is a serious life-threatening medical condition that occurs when the blood supply to part of the brain is cut off. Every year more than 15 million people worldwide have a stroke. Of these, one-third die, another one-third are left disabled/cognitive impairment and one-third recover 7 Post-stroke cognitive impairment Vascular cognitive impairment (VCI) is caused by damage to brain tissue, which occurs because of reduced supply of oxygen to the brain VCI is associated with both disability and survival Post-stroke cognitive impairment rates vary between studies: 10% patients had dementia before first stroke 10% develop dementia after first stroke And more than 1/3 develop cognitive impairment after recurrent stroke 8 Collaboration 1: Prevalence study In recent study of South London Stroke Register we have found that about quarter of patients develop cognitive impairment after the first-ever stroke (Douiri et al. Stroke. 2013;44:138-145) 9 10 Statistics Adjusted rates and 95% Confidence intervals Regression analyses to estimate relative risk ratios and hazard ratios 11 Mechanism! 12 Gap ! Treatment strategies already exist but focused on reducing the risk of cardiovascular events. Lack of understanding about how to prevent or delay the progression of cognitive impairment (damage to memory and mental skills) and dementia following a stroke. 13 Objectives and study design Objectives: Investigate whether vascular secondary prevention drugs are associated with a protective effect of global cognitive function after stroke Design and Setting: Population-based register of first ever stroke, the South London Stroke Register SLSR (1995-2011, 4413 patients) Outcome measure: Global cognitive function (Abbreviated Mental Test or Mini-Mental State Exam). Statistical methods: Generalized estimating equation Poisson models with robust standard errors and an exchangeable correlation matrix (treatment as time-varying covariate). 14 South London Stroke Register Population based register of first in a lifetime stroke in N. Lambeth & N. Southwark since 1995. Southwark Lambeth Source population of 357308 , made up of 56% White, 25% Black (7% Black Caribbean, 14% Black African, and 4% Black Other), 4% mixed (white and black) and 15% of other ethnic Group (UK census). 15 Collaboration 2: Secondary prevention and VCI 16 Results: longitudinal analysis by stroke type Adjusted Prescribed drugs by stroke type RR (95% CI) Haemorrhagic stroke Antihypertensive (mono-therapy) 1.19 (0.75, 1.88) Antihypertensive (combination-therapy) 1.14 (0.72, 1.79) Ischaemic stroke with AF Antihypertensive (mono-therapy) 1.21 (0.96, 1.52) Antihypertensive (combination-therapy) 1.02 (0.77, 1.35) Anticoagulant 0.88 (0.69, 1.12) Lipid-lowering 0.75 (0.54, 1.03) Ischaemic stroke without AF Antihypertensive (mono-therapy) 0.79 (0.70, 0.90) Antihypertensive (combination-therapy) 0.76 (0.66, 0.87) Antiplatelet (mono-therapy) 1.04 (0.94, 1.15) Antiplatelet (combination-therapy) 0.91 (0.79, 1.06) Lipid-lowering 0.88 (0.77, 1.00) Optimal therapy 0.55 (0.40, 0.77) 0.5.75 1 1.5 17 Results: Drug class Adjusted Prescribed drugs by stroke type RR (95% CI) Haemorrhagic stroke Diuretic 0.89 (0.51, 1.56) Angiotensin-converting enzyme inhibitors 1.32 (0.72, 2.44) Beta-blocker 2.29 (1.05, 4.99) Beta-blocker+Diuretic 0.78 (0.39, 1.58) Angiotensin-converting enzyme inhibitors+Diuretic 1.38 (0.88, 2.19) Ischaemic stroke with AF Diuretic 1.28 (0.94, 1.73) Angiotensin-converting enzyme inhibitors 1.16 (0.74, 1.82) Beta-blocker 1.24 (0.76, 2.03) Beta-blocker+Diuretic 1.02 (0.52, 2.00) Angiotensin-converting enzyme inhibitors+Diuretic 1.22 (0.81, 1.83) Warfarin 0.81 (0.59, 1.12) Statin 0.70 (0.44, 1.11) Ischaemic stroke without AF Diuretic 0.68 (0.57, 0.82) Angiotensin-converting enzyme inhibitors 0.79 (0.64, 0.98) Beta-blocker 0.90 (0.68, 1.18) Beta-blocker+Diuretic 0.81 (0.63, 1.04) Angiotensin-converting enzyme inhibitors+Diuretic 0.67 (0.55, 0.81) Aspirin 1.03 (0.92, 1.15) Aspirin+Dipyridamole 0.83 (0.68, 1.01) Statin 0.90 (0.76, 1.06) 0.5.75 1 1.5 18 Clinical Perspective Secondary prevention measures reduce not only the risk of recurrent ischemic episodes but also the risk of global cognitive decline. Treating with antihypertensive medication or a combination of aspirin and dipyridamole significantly protects against cognitive decline. The combination of antihypertensives, antithrombotics, and lipid- lowering drugs reduces the risk of cognitive impairment by about half. The protective effect was not seen in hemorrhagic stroke patients or in patients whose stroke was associated with atrial fibrillation. 19 Conclusion Different therapeutic strategies had different effects, suggesting that an optimal, personalised combination of antithrombotic, antihypertensive and lipid-lowering agents may protect patients from cognitive impairment. Controlling vascular risk factors after stroke helps preserve cognitive function Statin was associated with a lower risk of cognitive impairment, recurrence and survival. 20 Future work Internationally, STROKOG is a consortium of longitudinal studies of cognitive disorders following stroke, TIA or small vessel disease. Understanding the mechanism of post- stroke cognitive impairment. Management/monitoring Locally, linkage with primary care and mental health services Link with depression and social isolation Big data… Risk prediction (small vessel disease) CVD managements and stroke: primary/secondary prevention 21