VItamin D as a Novel Determinant of Injurious Cardiac Remodelling After acuTE Myocardial Infarction

Participants Patients who have recently (<72 hours) suffered an acute myocardial infarction will be invited to participate. The immediate post-infarction period is one of particular anxiety for patients and their carers and the project has been designed following advice from Patient-Public-Involvement advisers and also after discussion of the project at the Cardiovascular Clinical Research Facility PPI group meeting. Patient attendance will be kept to a minimum and dovetail with clinical investigations wherever possible. In view of the exploratory nature of the project, and the primary endpoint: left ventricular remodelling, exclusion and inclusion criteria have been limited, preferring instead to capture a real-world pragmatic view of changes in cardiac function following myocardial infarction. This is also important when viewed in context of the long-term plan to investigate the effects of vitamin D in ameliorating these adverse changes following a heart attack.

Data collection, administration and imaging protocol Demographic and clinical variables will be recorded on a bespoke database, which will also store biochemical information including vitamin D, PTH and vitamin D binding protein levels measured in routine blood samples. EDTA blood we will also be stored for analysis of vitamin D binding protein haplotypes. It is thought that vitamin D does not exhibit an acute phase response so levels are stable following a myocardial infarction. Each participant will undergo full bedside echocardiography within 72 hours of an acute ST elevation myocardial infarction, the results of which will be available for clinical use. European and American guidelines recommend echocardiography prior to discharge for patients admitted with a STEMI, so this scan will not be an extra procedure. The consent process will describe that the scan is clinically indicated but will also be used for research purposes. As such this proposed study will provide support for the clinical service.

VINDICATE (vitamin D replenishment in patients with chronic heart failure) examined remodelling and vitamin D over 12 months. Remodelling is expected to be accelerated in patients experiencing an acute STEMI (a recent study of acute iron supplementation compared scans at baseline and at three months) and indeed previous work describing the outcomes and predictors of remodelling in 284 STEMI patients did so using repeat echocardiography at 6 months.

Full conventional echocardiography will be carried out with grey-scale and tissue Doppler images recorded in the two and four chamber views using harmonics to improve border definition if necessary. All images will be stored in the 'echopac' digital imaging system and analysed offline. This analysis will include a calculation of LV end diastolic and end systolic volumes using the real-time three dimensional data and two-dimensional biplane discs (modified Simpson's) method by tracing the endocardial border excluding the papillary muscles. An average of three measurements will be used in the final analysis. The frame at the R-wave will be taken as end diastole, and the frame with the smallest LV cavity, as end systole. The LV end-systolic volume index (LVESVi) will be calculated as LVESV/body surface area.

Statistical considerations Study design and analysis plan have been developed with the help of Dr David Cairns, Principal Statistician at the Leeds Clinical Trials Research Unit.

Statistical analysis plan Clinical and technical variables known to be of relevance to outcomes in patients following myocardial infarction including pre-existing left ventricular function, anaemia, age, female sex, diabetes mellitus, previous chronic kidney disease will be collected.

Change in left ventricular volume is a reliable surrogate of change in outcome. Two possible markers of progressive remodelling will be analysed. The first will be to use the cut-off of 5% reduction in LVEF or a 15% increase in left ventricular end systolic volume index (LVESVi), two variables which have been frequently used in the literature as a measure of adverse remodeling. How changes in LVESVi and LVEF as continuous variables relate to the baseline clinical data will also be explored. In this analysis we will follow the guidelines described in STARD and TRIPOD statements to produce transparent and reproducible evidence which relate clinical variables to markers of progressive remodeling.

Hence, the initial approach will be to perform univariate tests of association (two sample t-tests, one way analysis of variance and chi-squared tests) to look at the relationship between the predictor variables of rhythm, past medical history (of ischaemic heart disease and diabetes mellitus), pre-existing left ventricular function, location of myocardial infarction, anaemia, age, female sex, previous chronic kidney disease and markers of progressive remodelling. Around one quarter (25%) of all returning patients are expected to fulfil the binary cut offs of >5% reduction in LVEF or >15% change in LVESVi. This will then be further developed to construct robust multivariable predictive models based on a step-wise logistic regression modelling procedure using bootstrap resampling to develop and internally validate the model. This approach will require sufficient patients' data to reliably allow regression models including all potentially relevant clinical variables.

Sample size prediction It has been assumed that for each independent predictor under investigation, there will need to be 10 'events' (>5% reduction in LVEF). With nine predictor variables, allowing for drop outs, withdrawal of consent or poor imaging, while recognising that the predictor variables are not likely to be independent of one another, it has been conservatively estimated that the study will need 90-100 subjects with 'events' (in this case remodelling) to robustly estimate a model. Since conservatively estimates suggest that around 25% of patients will experience remodelling, 400 patients will be scanned at baseline and 6 months.

However, around half of all patients presenting to Leeds with a STEMI come from other districts and may not want to return to LTHT for an echocardiogram. It has therefore been conservatively estimated that around 800 patients will need a scan to achieve 400 attendances at follow-up. Leeds performs primary angioplasty (for STEMI) on around 1200 patients per year such that allowing for weekends and refusals, there will be sufficient patients to complete enrolment within two years.

Follow-up and outcomes recording Outcomes data at one year, two years and five years following the index event will be collected using routinely collected hospital admissions and mortality data.