Journal of Rehabilitation Medicine 51-9 | Page 24

648 H. Ihle-Hansen et al. in Table II. The results from complete case analyses showed the same trends (data not shown). Unadjusted regression analysis applied to the inter- vention arm showed a significant association between increasing adherence to the intervention and improve- ment on TMT A, TMT B and MMSE. However, only MMSE remained significant in the adjusted analysis (B = 0.030 (95% CI 0.005–0.055, p = 0.020) (Table III). Table I. Baseline characteristics (n  = 362) Assessment Demographics Female, n (%) Age, years, mean (SD) Stroke subtype, n (%) Cerebral infarction Cerebral haemorrhage Risk factors, n (%) Hypertension Previous stroke/TIA Diabetes Coronary heart disease Atrial fibrillation Lung disease Assessments NIHSS, mean (SD) NIHSS < 8, n (%) NIHSS 8–16, n (%) NIHSS > 16, n (%) mRS, mean (SD) mRS 0–2, n (%) mRS 3–4, n (%) Intervention (n  = 177) Control (n  = 185) 78 (44.1) 71.4 (11.3) 65 (35.1) 72.0 (11.3) 163 (92.1) 14 (7.9) 166 (89.7) 19 (10.3) 85 26 24 18 28 18 104 (56.2) 35 (18.9)/16 (8.6) 27 (14.6) 24 (13.0) 41 (22.2) 23 (12.4) (48.0) (14.7)/20 (11.3) (13.6) (10.2) (15.8) (10.2) 1.47 (2.2) 172 (97.2) 5 (2.8) 0 (0) 1.44 (1.1) 141 (79.7) 36 (20.3) DISCUSSION 1.70 (2.5) 179 (96.8) 6 (3.2) 0 (0) 1.44 (1.1) 146 (78.9) 39 (21.1) Coronary heart disease: previous myocardial infarction or present angina pectoris and heart failure; atrial fibrillation: permanent or paroxysmal atrial fibrillation; NIHSS: National Institute of Health Stroke Scale; IQR: interquartile range; mRS: modified Rankin Scale; TIA: transient ischaemic attack; SD: standard deviation. 186 patient were randomised to intervention and 194 to standard care (9). After exclusion of 18 patients who died during follow-up (9 from the control group and 9 from the intervention group), 362 patients were included in these ITT analyses; 185 were assigned to control and 177 to intervention. Mean age was 71.7 years (SD 11.3), 39.5% were female and 329 (90.9%) had had ischaemic stroke. The baseline characteristics are shown in Table I. Measures of cognitive function and emotional symp- toms showed a slight decline in both groups during follow-up (Table II). The adjusted mean difference between groups for TMT A was 8.54 (95% CI 0.7 to 16.3), p = 0.032, for TMT B 8.6 (95% CI –16.5 to 33.6), p = 0.50, for MMSE –0.1 (95% CI –0.8 to 0.6), p = 0.77, for HADS A –0.2 (95% CI –0.9 to 0.5), p = 0.456 and for HADS D –0.1 (95% CI –0.7 to 0.5), p = 0.76). The results based on multiple imputations are presented This study investigated the effectiveness of an indi- vidualized physical activity and exercise intervention programme on cognitive and emotional function post- stroke compared with usual care. The intervention involved regular coaching to perform 30 min physical activity daily every day and 45–60 min of physical ex- ercise with 2–3 bouts of vigorous intensity levels every week. No clinically relevant effect of this programme was found on cognitive or emotional functioning after 18 months compared with usual care. However, regres- sion analysis applied to the treatment arm only showed a positive association between increasing adherence to the intervention and cognitive function. To our knowledge, this is the first RCT to investigate the effect of a long-term physical activity programme on cognitive and emotional function post-stroke in a rather large cohort with high mean age. The lack of clinical impact on cognitive measures is in line with the neutral results shown by the primary and secondary functional measures (9). There was no effect on emo- tional function, which is in contrast to a recent study published by Graven et al. (18), who also included life-management, barriers and social network. How­ ever, our explorative analysis showed an association between good adherence to the intervention and better cognitive function, in line with the findings from the meta-analysis (7). Since motor recovery after stroke tends to reach a plateau phase after 10 weeks (19), the initial improve- Table II. Results of the intervention. Analyses based on multiple imputation of missing values Intervention group (n=177) Control group (n=185) Between-group differences, change from baseline to 18 months* Outcome Baseline Mean (SE) [n] 18 months Mean (SE) [n] Baseline Mean (SE) [n] 18 months Mean (SE) [n] Estimate (95% CI) p-value TMT A TMT B MMSE HADS A HADS D 60.5 (2.9) [156] 149.4 (7.4) [132] 27.9 (0.2) [176] 3.6 (0.3) [177] 2.9 (0.2) [177] 70.1 (4.2) [133] 155.8 (9.7) [108] 27.5 (0.3) [143] 3.6 (0.3) [149] 3.7 (0.2) [149] 64.4 (2.6) [169] 169.8 (7.9) [142] 28.0 (0.2) [182] 3.7 (0.3) [180] 3.5 (0.2) [180] 66.0 (3.3) [139] 161.3 (12.0) [101] 27.5 (0.3) [156] 3.8 (0.3) [157] 3.9 (0.3) [157] 8.54 (0.74 to 16.3) 8.6 (–16.5 to 33.6) –0.1 (–0.8 to 0.6) –0.2 (–0.9 to 0.5) –0.1 (–0.7 to 0.5) 0.032 0.50 0.77 0.56 0.76 *Regression coefficient for treatment group in regression analysis, adjusted for age, sex, stroke severity (modified Rankin scale at inclusion), hospital site and the baseline measure of the relevant variable (TMT A, TMT B, MMS, HADS A or HADS B). SE: standard error; CI: confidence interval; MMSE: Mini Mental State Examination; TMT A: Trail Making Test A; TMT B: Trail Making Test B; HADS: Hospital Anxiety and Depression Scale; HADS A: HADS Anxiety; HADS D: HADS Depression. www.medicaljournals.se/jrm