Journal of Rehabilitation Medicine 51-9 | Page 23

Physical activity post-stroke to prevent cognitive impairment emotional function in stroke survivors. In addition, this study explored the association between adherence to the intervention and the outcome variables. METHODS LAST is a prospective, randomized, single-blinded, multicentre, 2-arm parallel group clinical trial. Between October 2011 and January 2016, patients were recruited 3 months post-stroke from 2 Norwegian hospitals; Trondheim University Hospital and Bærum Hospital with intervention and follow-up for 18 months in the primary healthcare services in the municipalities of Trondheim, Asker and Bærum (NCT01467206) (10). The current study analysed secondary outcomes. Participants Inclusion criteria were: patients aged over 18 years, with first- ever or recurrent stroke due to infarction and intracerebral haemorrhage, community-dwelling with modified Rankin Scale (mRS) score < 5. Exclusion criteria were serious medical co- morbidity with short life-expectancy, cognitive deficits, as eva- luated by Mini Mental State Examination (MMSE) < 21 points (or < 17 points for patients with aphasia) and contraindication to participate in motor training (10). Neurological impairment was measured by the National Institute of Health Stroke Scale (NIHSS), which ranges from zero to 42. The patients were trea- ted in the acute and subacute phase according to international and national guidelines from 2011 (11). Risk assessment and individualized secondary prevention was initiated at discharge from hospital after the index stroke, with recommendation for long-term risk factor control as part of usual care. Randomization Participants were stratified according to functional level post- stroke (mRS > 2 points), age ≥ 80 years and recruitment site, and randomized by a web-based randomization system developed and administered by the Unit of Applied Clinical Research, Faculty of Medicine and Health Sciences, Norwegian Univer- sity of Science and Technology, Trondheim, Norway. Patients randomized to the control group received usual care. Intervention In addition to usual care, the intervention group received regular individualized coaching performed by physiotherapists, aiming to achieve physical activity 30 min daily, and 45–60 min physical exercise including 2–3 bouts of vigorous activity every week. The intervention did not include supervised or observed training. The physiotherapists carried out home visits to provide education and tailored training programmes, based on the patient’s preferences and goals. The patients were encouraged to report their activi- ties in standardized training diaries. The patients had a monthly meeting with their coach, going through training diaries, with reassessment according to needs. These meetings were face-to- face in the first 6 months. For the next 6 months, every second meeting could take place as a phone meeting. During the final 6 months, 4 of the 6 meetings could take place as a phone meeting. Outcomes Cognitive function was measured using standardized neuropsy- chological tests assessing for global function, processing speed 647 and executive function (normal scores); MMSE (minimum score 0 to maximum score 30), Trail Making A and B (TMT A (< 60 s considered a normal score) and B (< 180 s considered a normal score) (12, 13) at inclusion 3 months post-stroke and at 18 months follow-up. Symptoms of anxiety and depression were assessed by Hospital Anxiety and Depression Scale (HADS A and D) (14). Each domain consists of 7 questions with scores ranging from 0 (no symptoms) to 3 points. The outcomes were changes in cognitive and emotional measures from baseline to 18 months follow-up. Adherence to the intervention was measured using self- reported activity through the diaries and evaluation by the physiotherapists, defined as the number of weeks patients in the intervention group completed at least 210 min of physical activity (minimum 30 min every day) and 45 min of exercise (in total per week). Despite considerable variation, it has been shown that the participants established and maintained mode- rate-to-good adherence to the intervention (15). Statistical analyses Sample size was calculated according to the primary outcome (Motor Assessment Scale) in the main study (9). Descriptive statistics are given as mean (SD), median (interquartile range), or frequency (percentage), where relevant. According to the CONSORT instructions, all outcomes underwent intention-to- treat (ITT) analysis (16). Linear regression analyses were used, with TMT A and B, MMSE and HADS A and D, respectively, as dependent variables, and intervention group, age, sex, stroke severity (measured by mRS), hospital site, treatment group, and cognitive or emotional status at inclusion as covariates. For patients in the treatment arm, linear regression analyses were carried out with the cognitive and emotional outcome measures, TMT A, TMT B, MMSE, HADS A, and HADS D, at 18-month follow-up from baseline, as dependent variables, one at a time, with adherence to physical activity and exercise as primary covariate. These regression analyses were carried out both unadjusted and adjusted for the following covariates, one at a time and simultaneously: age, sex, mRS, MMSE, if appropriate, and the corresponding outcome variable score, all measured at baseline 3 months post-stroke. On HADS and MMSE, the subjects with less than 50% missing values on a scale, the missing values were imputed singly using the expectation–maximization algorithm on that scale. Thereafter, multiple imputation was used to impute missing scale sums and other missing values, with m = 100 imputation, as recommended by van Buuren (17). Patients who died during follow-up were ex- cluded from the analyses. The normality of residuals was checked by visual inspection of QQ plots. Two-sided p-values < 0.05 were considered statistically significant, and 95% confidence intervals (95% CI) are reported where relevant. Ethics The study was approved by the Regional Committee of Medical and Health Research Ethics (REK no 2011/1427) and registered with ClinicalTrias.gov, number NCT01467206. All participants provided informed consent. RESULTS In total, 380 patients were recruited and randomized in the main study (the trial profile is reported elsewhere; J Rehabil Med 51, 2019