MRC/BHF HEART PROTECTION STUDY

The Medical Research Council/British Heart Foundation Heart Protection Study (HPS) is led by the Nuffield Department of Population Health (NDPH) at the University of Oxford. The University of Oxford is sole data controller for this Agreement.

This Agreement will permit the use of the data already provided for the (MRC/BHF) Heart Protection Study (HPS) cohort for long term follow up.

The MRC/BHF Heart Protection Study was designed to be really large, to involve a substantial reduction in blood LDL cholesterol levels and a substantial increase in antioxidant vitamin levels, and to include a wide range of patients at substantial risk of death from heart disease and other causes.

The Heart Protection Study (HPS) was a large randomised controlled trial. Between 1994 and 1997, 20,536 individuals in the UK at increased risk of coronary heart disease were randomised to 40mg simvastatin daily versus matching placebo, and (in a 2X2 factorial design) to anti-oxidant vitamin supplementation with vitamins E, C and beta-carotene versus placebo. Participants took trial medications for an average of 5-years (scheduled treatment period), and the main trial closed in 2001. The aim was to study the overall effects on survival by preventing heart attacks, strokes and other major vascular events.

The main results of HPS (published in 2002) clearly showed that allocation to cholesterol-lowering therapy with simvastatin was associated with a 20% reduction in the risk of heart attacks and stroke, and a reduction in all-cause mortality, chiefly driven by a very significant reduction in vascular deaths. These results changed clinical practice worldwide, and statins are now prescribed widely to many millions of patients.

Participants were randomised into HPS from July 1994 to May 1997. After randomisation finished in 1997, participants continued on randomised treatment and were followed-up in clinics until 2001. All participants stopped taking treatment in 2001. Following this, annual questionnaires were mailed to the surviving participants between 2002 and 2007. This follow-up was supplemented with cause-specific mortality data provided by the Office for National Statistics (ONS) and incident cancers via national cancer registries. Additionally, and with the necessary approvals, Hospital Episodes Statistics (HES) data were provided by the NHS Information Centre (now called NHS England), giving details on non-fatal events occurring in the study population.

HPS has demonstrated that cholesterol-lowering with 40mg simvastatin daily for an average of 5-years substantially reduces the risk of major vascular events in a much wider range of individuals than previously thought to benefit from such treatment. The reductions in risk increased throughout the scheduled treatment period, suggesting that more prolonged treatment might result in larger benefits. No adverse effects on non-vascular mortality or cancer were observed during the statin treatment period. HPS also assessed the effects of antioxidant vitamin supplementation (vitamins E, C and beta-carotene), with no significant benefits or hazards observed during the 5-year treatment period. Prolonged follow-up of participants via routinely collected health records and central registries allows reliable assessment of any delayed effects on major outcomes.

NDPH has a strong research interest in the reliable assessment of the effects of LDL cholesterol lowering and the HPS investigators wish to continue long-term follow-up of this cohort, which will add important data about the effects of LDL cholesterol lowering on important chronic health outcomes. In addition to the effects of randomly allocated treatment, the study team will look at the effect of clinical, genetic, and bio-marker data on future health.

The study team have three similar studies with common aims but for which three separate Data Sharing Agreements are held with NHS England. Due to the similarities between the studies and the cohorts used, where any one study (this Agreement, in addition DARS-NIC-148341-TC6TD (SEARCH) and DARS-NIC-147885-0TV66 (HPS2-THRIVE)) does not give sufficient power for an analysis, the study team will perform meta-analyses to a common protocol from each study. Where the randomised allocations are similar, the study team will perform study level and individual participant data meta-analyses to look at the effect of variables on major health events such as stroke, Myocardial Infarction (MI) & dementia.

The first planned analyses will be based on at least 20 years’ follow-up from trial initiation with further analyses planned at approximately 5 yearly intervals based on on-going linkage to NHS records.

Cholesterol levels are of particular interest in relation to neuro-degenerative diseases because a genetic risk factor for Alzheimer’s disease is the ApoE gene, which binds to receptors in low-density lipo-protein, and is involved in the neuronal transport of cholesterol in the brain. People with this gene have higher levels of total cholesterol and triglycerides than those without, and therefore higher blood LDL cholesterol is one potential mechanism for the deleterious effect of the gene on cognitive function.

In observational studies, there is clearly a strong link between the ApoE gene and dementia. Higher midlife total cholesterol is associated with later life cognitive impairment or dementia, though the magnitude of this effect or the extent to which this is mediated by confounding by other vascular risk factors is unclear. Dementia is a condition that develops over a long period before manifesting in a clinical diagnosis. In the short-term (up to 10-years) lower cardiovascular risk factor levels have often been associated with an increased risk of dementia, which may be because of reverse causal effects of the incipient dementia leading to lower levels. However, raised mid-life levels of cardiovascular risk factors (such as LDL-cholesterol) have been found to be associated with increased risk of dementia 15-20 years later. There is little data, however, on whether raised levels of cardiovascular risk factors at older ages are associated with an increased risk of dementia 15-20 years later. Continued follow-up for dementia in studies (such as HPS) in older people initiated many years ago is therefore extremely valuable for investigating such effects now.

HPS recruited 20,536 UK participants with high vascular disease risk and with a mean age of about 64 at randomisation.

Dementia is a leading cause of death in the UK and it is likely that over a third of this population will develop dementia at some point in their lifetime. Hence, many people in HPS may by now have developed dementia. This study now constitutes a uniquely rich resource for study of the relationships of vascular risk factors to dementia incidence many years later. This study will look at the association of vascular risk factors measured at baseline with dementia incidence at various times into the future, with longer delays between measurement of risk factors at recruitment and incidence of dementia being particularly valuable.

The study team have previously received data from NHS England. Under this Agreement further data are requested and the data controller wish to retain all data previously disseminated. Additionally, the data controller now have permission, under HRA Section 251 support, to request hospital episodes and mental health data for the cohort. These data will enable the data controller to examine the effect of baseline clinical, genetic and biomarker variables on major health events, including major vascular events (including, but not limited to strokes of different types and myocardial infarction), cancers, renal disease and dementia. The data controller will ascertain these events chiefly though ICD-10 coding. The study team have found in previous work that the majority of cases of chronic disease (including dementia) can be found in admitted patient care, death and mental health records.

The purpose of previously requesting data under this agreement, was to determine factors that contribute to the health of trial participants in the longer term.

The principle research objectives are to directly assess the very long-term effects of both:

i) Around 5-years of statin treatment (40mg) versus matching placebo, and

ii) Around 5-years use of antioxidant vitamin supplements (vitamin E, vitamin C, and beta-carotene) versus matching placebo,

on major health events (i.e. major vascular events, cancer, dementia) and death.

Secondary research objectives are to investigate associations between both various patient characteristics (e.g. age, gender, prior disease, blood pressure, height and weight) and also blood test results (including genetic analyses), and the risk of developing important medical conditions (e.g. heart attacks, strokes, cancers, and dementia) in later life.

Tertiary research objectives will include assessment during extended follow-up of the effects on other major adverse outcomes (e.g. site-specific cancer, cerebral haemorrhage, hospitalisation for angina and for various other causes).

In order to determine the occurrence of the diagnosis of particular illnesses, the study team have linked trial participants to hospitalisation, mental health, death and cancer records. The study team previously requested linkage to Hospital Episode Statistics (HES) for Admitted Patient Care, the Mental Health datasets, Civil Registration – Deaths, Cancer Registrations, & Demographics.

The study team requested linkage of individual participants by their identifying information (name, NHS number, date of birth, gender) by NHS England. NHS England returned data to the Nuffield Department of Population Health at the University of Oxford with each participant's trial identity number with data from the linkage, i.e. pseudonymised record-level data.

Although the University of Oxford has not asked for personal identifiers under this Agreement, the remaining supplied data was still linked to the existing long-term follow-up study database, which does store the identifying data. There is an ongoing need to retain identifying data to enable further data linkages, such as that which will take place under this Data Sharing Agreement.

Direct identifiers will need to be retained whilst linkages are made between datasets before all data are identified primarily with the study ID, in order to pseudonymise the long-term follow-up dataset. The intention is to keep a pseudonymised dataset for future analysis for as long as is necessary, and this will be reviewed on a yearly basis.

In addition, it is NDPH policy to retain research data for at least 25-years after the end of a the trial as per the 2014 Clinical Trials Regulation (EU) No 536/2014: https://ec.europa.eu/health/system/files/2016-11/reg_2014_536_en_0.pdf

The data subjects will all be participants of the original randomized controlled trial. The study team will obtain data from each participant from the time they began to take part in the trial (recruitment from July 1994 - May 1997) to the date of linkage. This allows researchers to:

i) compare the occurrence of events recorded by the study during follow-up with events recorded by electronic health records;

ii) compare the very long term incidence of major disease by randomised treatments and baseline factors.

Further data releases will allow the study team to continue follow-up of participants for their entire lifespan including cause of death.

Participants were resident in and recruited from across the UK but may have moved between nations since recruitment, under this Agreement the study team would like to link to data held across England and Wales.

Linking participants to data held by NHS England is the least intrusive way of achieving the study's purpose. The data controller believes that re-approaching participants for further consent would lead to potentially very large biases in ascertainment of major health conditions (because of non-responder biases particularly amongst those in poorest health, or those who have died) and potentially lead to distress. Therefore, the data controller has section 251 support from the Health Research Authority (HRA) Confidentiality Advisory Group (CAG) for this linkage (ref: 20/CAG/0113). CAG has confirmed that the national data opt out is applied at the point of disclosure and does not apply retrospectively.

The data under this Agreement is treated as identifiable because the data controller continues to retain the patient identifiers which have been provided to NHS England.

The study team request linkage only to those data-sets that contain information on medical diagnoses, for the participants of the study, minimised to the date range of the post-trial follow-up, currently 1st July 1997 to 1st October 2035.

The study team will also link to equivalent datasets received from Digital Health and Care Wales (previously NHS Wales Informatics Services), Public Health Scotland and the NHS Central Register (NHSCR).

Blood samples collected during the main trial may be used for the genomic arm of this study (support has been received from South Central – Oxford B REC for this work). The genomic analyses will be undertaken at specialist laboratories. This will involve sending pure DNA samples. DNA is not a relevant material under the HTA and all samples will be sent with only the participant study identifier and no personal data. All data will be returned to Oxford for statistical analyses.

The study team intend to link data from participant DNA analyses to NHS England data. As demonstrated by the HPS long-term follow-up results, linkage to longer term outcomes (including those identified from NHS England data) is of considerable importance in fully understanding response to statins and cardiovascular risk and its consequences as a whole. Furthermore, genetic data may be able to add insights into our understanding of the effects of statins, determinants and mechanisms of risk factors and disease outcomes, that can contribute to the development of better patient treatments and precision medicine approaches.

The processing of the data under this Agreement will be carried out in accordance with Article 6(1)(e) as processing is necessary for a task carried out in the public interest. The aim of the study is to provide reliable evidence about the very long term effects of cholesterol-lowering treatments in the UK population and is therefore in the public interest.

Processing of the data is in accordance with Article 9(2)(j) exemption, i.e. that the processing of the data is necessary for scientific research. The scientific research aim of the study is to provide reliable evidence about the very long term effects of cholesterol-lowering treatments on important health outcomes.

The research team have taken the opinion of the South Central – Oxford B Research Ethics Committee, who have granted approval to the study in its current form. Furthermore, potential harm to the public will be limited by data pseudonymisation, data minimisation, and technical and organisational safeguards. Analysis will occur in a safe NHS Data Security and Protection (DSP) Toolkit compliant environment and identifiers will not be included in the analysis. Although the study is not still in direct contact with participants, at the time the study team asked for the opinion of available participants in these and similar studies, they agreed that the use of the data in the way that the study team have proposed is reasonable.

All processing of NHS England data will take place within the University of Oxford, therefore Oxford are listed as sole data processor under this Agreement.

One of the Chief Investigators is employed by both Gloucestershire Hospitals NHS Foundation Trust and the University of Oxford. This Chief Investigator will analyse the data under this Agreement solely in their role as a substantive employee of the University of Oxford, and therefore Cheltenham General Hospital is not listed as a data controller under this Agreement.

Funding for this study is provided internally from the University of Oxford.

The primary outputs from this data will be academic, and will include submissions to peer reviewed journals such as New England Journal of Medicine and the Lancet, and conferences such as the American Heart Association meeting.

The study team will only present data at an aggregated level with small numbers suppressed. The study team will present actual and modelled data in graphical and tabular format.

The NDPH contributes widely to national and international health policy, particularly in the area of vascular risk prevention. It contributes to debate with academic papers, conference participation, lectures to the public and advice to government (including NHS England). Examples of the impact of the work performed by NDPH up to 2014 is available from: https://results.ref.ac.uk/(S(ep5gbndxsprqnc0kork3mjyu))/Submissions/Impact/728

The study team will share outputs via the following channels:

- Study website

- Open lectures and talks

- Exhibition at public events

- Posters

- Press/media engagement and other public promotion of the research

The data controller aims to issue the next publication by the end of 2023.

All outputs will only contain results in highly aggregated format and as statistical summaries and measures of association. Small numbers will be suppressed in line with the HES Analysis Guide. Record level information will not be released to any third party.

Participants have been provided with updates throughout the years since the study began in 1994. Newsletters were mailed to participants between 1996 and 2000. In 2001, a letter was mailed which included information about the main trial results and details of the HPS study website. This website has been regularly updated since 2001 and in 2020 the study team uploaded two new documents to the study, one was a lay summary for the HPS follow-up study and one was a privacy notice for the follow-up study.

Outputs for the HPS study so far include multiple publications in peer-reviewed journals and presentations at international conferences. Information about the publications can be found on the HPS website https://www.ctsu.ox.ac.uk/research/hps

This study has strongly influenced the labelling of statin medication internationally, treatment guidelines, and the resulting changes in prescribing have contributed to reductions in mortality and morbidity from heart attack and blocked arteries (‘ischaemic stroke’) in many countries.

The main results of HPS (published in 2002) clearly showed that allocation to cholesterol-lowering therapy with simvastatin was associated with a 20% reduction in the risk of heart attacks and stroke, and a reduction in all-cause mortality, chiefly driven by a very significant reduction in vascular deaths. These results changed clinical practice worldwide, and statins are now prescribed widely to many millions of patients. The following outcomes for the study influenced NICE Guidelines within the UK:

1) Statin therapy significantly reduces the incidence of coronary and other major vascular events to a similar extent, irrespective of KIF6 genotype. Consequently, the use of KIF6 genotyping to guide statin therapy is not warranted.

2) These findings provide considerable reassurance that lowering total cholesterol concentrations by more than 1 mmol/L for an average of 5 years does not produce adverse effects on non-vascular mortality or cancer incidence. Moreover, among the many different types of high-risk individual studied, simvastatin 40 mg daily consistently produced substantial reductions in vascular (and, hence, all-cause) mortality, as well as in the rates of non-fatal heart attacks, strokes and revascularisation procedures.

3) Treatment with statins is cost effective in a wider population than is routinely treated at present.

4) Adding simvastatin to existing treatments safely produces substantial additional benefits for a wide range of high-risk patients, irrespective of their initial cholesterol concentrations. Allocation to 40 mg simvastatin daily reduced the rates of myocardial infarction, of stroke, and of revascularisation by about one-quarter. After making allowance for non-compliance, actual use of this regimen would probably reduce these rates by about one-third. Hence, among the many types of high-risk individual studied, 5 years of simvastatin would prevent about 70-100 people per 1000 from suffering at least one of these major vascular events (and longer treatment should produce further benefit). The size of the 5-year benefit depends chiefly on such individuals' overall risk of major vascular events, rather than on their blood lipid concentrations alone.

5) The present study provides direct evidence that cholesterol-lowering therapy is beneficial for people with diabetes even if they do not already have manifest coronary disease or high cholesterol concentrations. Allocation to 40 mg simvastatin daily reduced the rate of first major vascular events by about a quarter in a wide range of diabetic patients studied. After making allowance for non-compliance, actual use of this statin regimen would probably reduce these rates by about a third. For example, among the type of diabetic patient studied without occlusive arterial disease, 5 years of treatment would be expected to prevent about 45 people per 1000 from having at least one major vascular event (and, among these 45 people, to prevent about 70 first or subsequent events during this treatment period). Statin therapy should now be considered routinely for all diabetic patients at sufficiently high risk of major vascular events, irrespective of their initial cholesterol concentrations.

6) Much larger numbers of people in the present study suffered a stroke than in any previous cholesterol-lowering trial. The results demonstrate that statin therapy rapidly reduces the incidence not only of coronary events but also of ischaemic strokes, with no apparent effect on cerebral haemorrhage, even among individuals who do not have high cholesterol concentrations. Allocation to 40 mg simvastatin daily reduced the rate of ischaemic strokes by about one-quarter and so, after making allowance for non-compliance in the trial, actual use of this regimen would probably reduce the stroke rate by about a third. HPS also provides definitive evidence that statin therapy is beneficial for people with pre-existing cerebrovascular disease, even if they do not already have manifest coronary disease.

7) Research looking at survivors in the HPS study found that the combination of weight loss and high vascular risk identified individuals at high risk of dementia. These results suggest that weight loss could be considered when developing dementia risk scores.

The Medical Research Council/British Heart Foundation Heart Protection Study (HPS) is led by the Nuffield Department of Population Health (NDPH) at the University of Oxford. The University of Oxford is sole data controller for this Agreement.

This Agreement will permit the use of the data already provided for the (MRC/BHF) Heart Protection Study (HPS) cohort for long term follow up.

The MRC/BHF Heart Protection Study was designed to be really large, to involve a substantial reduction in blood LDL cholesterol levels and a substantial increase in antioxidant vitamin levels, and to include a wide range of patients at substantial risk of death from heart disease and other causes.

The Heart Protection Study (HPS) was a large randomised controlled trial. Between 1994 and 1997, 20,536 individuals in the UK at increased risk of coronary heart disease were randomised to 40mg simvastatin daily versus matching placebo, and (in a 2X2 factorial design) to anti-oxidant vitamin supplementation with vitamins E, C and beta-carotene versus placebo. Participants took trial medications for an average of 5-years (scheduled treatment period), and the main trial closed in 2001. The aim was to study the overall effects on survival by preventing heart attacks, strokes and other major vascular events.

The main results of HPS (published in 2002) clearly showed that allocation to cholesterol-lowering therapy with simvastatin was associated with a 20% reduction in the risk of heart attacks and stroke, and a reduction in all-cause mortality, chiefly driven by a very significant reduction in vascular deaths. These results changed clinical practice worldwide, and statins are now prescribed widely to many millions of patients.

Participants were randomised into HPS from July 1994 to May 1997. After randomisation finished in 1997, participants continued on randomised treatment and were followed-up in clinics until 2001. All participants stopped taking treatment in 2001. Following this, annual questionnaires were mailed to the surviving participants between 2002 and 2007. This follow-up was supplemented with cause-specific mortality data provided by the Office for National Statistics (ONS) and incident cancers via national cancer registries. Additionally, and with the necessary approvals, Hospital Episodes Statistics (HES) data were provided by the NHS Information Centre (now called NHS Digital), giving details on non-fatal events occurring in the study population.

HPS has demonstrated that cholesterol-lowering with 40mg simvastatin daily for an average of 5-years substantially reduces the risk of major vascular events in a much wider range of individuals than previously thought to benefit from such treatment. The reductions in risk increased throughout the scheduled treatment period, suggesting that more prolonged treatment might result in larger benefits. No adverse effects on non-vascular mortality or cancer were observed during the statin treatment period. HPS also assessed the effects of antioxidant vitamin supplementation (vitamins E, C and beta-carotene), with no significant benefits or hazards observed during the 5-year treatment period. Prolonged follow-up of participants via routinely collected health records and central registries allows reliable assessment of any delayed effects on major outcomes.

NDPH has a strong research interest in the reliable assessment of the effects of LDL cholesterol lowering and the HPS investigators wish to continue long-term follow-up of this cohort, which will add important data about the effects of LDL cholesterol lowering on important chronic health outcomes. In addition to the effects of randomly allocated treatment, the study team will look at the effect of clinical, genetic, and bio-marker data on future health.

The study team have three similar studies with common aims but for which three separate Data Sharing Agreements are held with NHS Digital. Due to the similarities between the studies and the cohorts used, where any one study (this Agreement, in addition DARS-NIC-148341-TC6TD (SEARCH) and DARS-NIC-147885-0TV66 (HPS2-THRIVE)) does not give sufficient power for an analysis, the study team will perform meta-analyses to a common protocol from each study. Where the randomised allocations are similar, the study team will perform study level and individual participant data meta-analyses to look at the effect of variables on major health events such as stroke, Myocardial Infarction (MI) & dementia.

The first planned analyses will be based on at least 20 years’ follow-up from trial initiation with further analyses planned at approximately 5 yearly intervals based on on-going linkage to NHS records.

Cholesterol levels are of particular interest in relation to neuro-degenerative diseases because a genetic risk factor for Alzheimer’s disease is the ApoE gene, which binds to receptors in low-density lipo-protein, and is involved in the neuronal transport of cholesterol in the brain. People with this gene have higher levels of total cholesterol and triglycerides than those without, and therefore higher blood LDL cholesterol is one potential mechanism for the deleterious effect of the gene on cognitive function.

In observational studies, there is clearly a strong link between the ApoE gene and dementia. Higher midlife total cholesterol is associated with later life cognitive impairment or dementia, though the magnitude of this effect or the extent to which this is mediated by confounding by other vascular risk factors is unclear. Dementia is a condition that develops over a long period before manifesting in a clinical diagnosis. In the short-term (up to 10-years) lower cardiovascular risk factor levels have often been associated with an increased risk of dementia, which may be because of reverse causal effects of the incipient dementia leading to lower levels. However, raised mid-life levels of cardiovascular risk factors (such as LDL-cholesterol) have been found to be associated with increased risk of dementia 15-20 years later. There is little data, however, on whether raised levels of cardiovascular risk factors at older ages are associated with an increased risk of dementia 15-20 years later. Continued follow-up for dementia in studies (such as HPS) in older people initiated many years ago is therefore extremely valuable for investigating such effects now.

HPS recruited 20,536 UK participants with high vascular disease risk and with a mean age of about 64 at randomisation.

Dementia is a leading cause of death in the UK and it is likely that over a third of this population will develop dementia at some point in their lifetime. Hence, many people in HPS may by now have developed dementia. This study now constitutes a uniquely rich resource for study of the relationships of vascular risk factors to dementia incidence many years later. This study will look at the association of vascular risk factors measured at baseline with dementia incidence at various times into the future, with longer delays between measurement of risk factors at recruitment and incidence of dementia being particularly valuable.

The study team have previously received data from NHS Digital. Under this Agreement further data are requested and the data controller wish to retain all data previously disseminated. Additionally, the data controller now have permission, under HRA Section 251 support, to request hospital episodes and mental health data for the cohort. These data will enable the data controller to examine the effect of baseline clinical, genetic and biomarker variables on major health events, including major vascular events (including, but not limited to strokes of different types and myocardial infarction), cancers, renal disease and dementia. The data controller will ascertain these events chiefly though ICD-10 coding. The study team have found in previous work that the majority of cases of chronic disease (including dementia) can be found in admitted patient care, death and mental health records.

The purpose of requesting the data under this Agreement is to determine factors that contribute to the health of trial participants in the longer term.

The principle research objectives are to directly assess the very long-term effects of both:

i) Around 5-years of statin treatment (40mg) versus matching placebo, and

ii) Around 5-years use of antioxidant vitamin supplements (vitamin E, vitamin C, and beta-carotene) versus matching placebo,

on major health events (i.e. major vascular events, cancer, dementia) and death.

Secondary research objectives are to investigate associations between both various patient characteristics (e.g. age, gender, prior disease, blood pressure, height and weight) and also blood test results (including genetic analyses), and the risk of developing important medical conditions (e.g. heart attacks, strokes, cancers, and dementia) in later life.

Tertiary research objectives will include assessment during extended follow-up of the effects on other major adverse outcomes (e.g. site-specific cancer, cerebral haemorrhage, hospitalisation for angina and for various other causes).

In order to determine the occurrence of the diagnosis of particular illnesses, the study team will need to link trial participants to hospitalisation, mental health, death and cancer records. The study team therefore request linkage to Hospital Episode Statistics (HES) for Admitted Patient Care, the Mental Health datasets, Civil Registration – Deaths, Cancer Registrations, & Demographics.

The study team request linkage of individual participants by their identifying information (name, NHS number, date of birth, gender) by NHS Digital. NHS Digital will return data to the Nuffield Department of Population Health at the University of Oxford with each participant's trial identity number with data from the linkage, i.e. pseudonymised record-level data.

Although the University of Oxford is not asking for personal identifiers under this Agreement, the remaining supplied data will still be linked to the existing long-term follow-up study database, which does store the identifying data. There is an ongoing need to retain identifying data to enable further data linkages, such as that which will take place under this Data Sharing Agreement.

Direct identifiers will need to be retained whilst linkages are made between datasets before all data are identified primarily with the study ID, in order to pseudonymise the long-term follow-up dataset. The intention is to keep a pseudonymised dataset for future analysis for as long as is necessary, and this will be reviewed on a yearly basis.

In addition, it is NDPH policy to retain research data for at least 25-years after the end of a the trial as per the 2014 Clinical Trials Regulation (EU) No 536/2014: https://ec.europa.eu/health/system/files/2016-11/reg_2014_536_en_0.pdf

The data subjects will all be participants of the original randomized controlled trial. The study team will obtain data from each participant from the time they began to take part in the trial (recruitment from July 1994 - May 1997) to the date of linkage. This allows researchers to:

i) compare the occurrence of events recorded by the study during follow-up with events recorded by electronic health records;

ii) compare the very long term incidence of major disease by randomised treatments and baseline factors.

Further data releases will allow the study team to continue follow-up of participants for their entire lifespan including cause of death.

Participants were resident in and recruited from across the UK but may have moved between nations since recruitment, under this Agreement the study team would like to link to data held across England and Wales.

Linking participants to data held by NHS Digital is the least intrusive way of achieving the study's purpose. The data controller believes that re-approaching participants for further consent would lead to potentially very large biases in ascertainment of major health conditions (because of non-responder biases particularly amongst those in poorest health, or those who have died) and potentially lead to distress. Therefore, the data controller has section 251 support from the Health Research Authority (HRA) Confidentiality Advisory Group (CAG) for this linkage (ref: 20/CAG/0113). CAG has confirmed that the national data opt out is applied at the point of disclosure and does not apply retrospectively.

The data under this Agreement is treated as identifiable because the data controller continues to retain the patient identifiers which have been provided to NHS Digital.

The study team request linkage only to those data-sets that contain information on medical diagnoses, for the participants of the study, minimised to the date range of the post-trial follow-up, currently 1st July 1997 to 1st October 2035.

The study team will also link to equivalent datasets received from Digital Health and Care Wales (previously NHS Wales Informatics Services), Public Health Scotland and the NHS Central Register (NHSCR).

Blood samples collected during the main trial may be used for the genomic arm of this study (support has been received from South Central – Oxford B REC for this work). The genomic analyses will be undertaken at specialist laboratories. This will involve sending pure DNA samples. DNA is not a relevant material under the HTA and all samples will be sent with only the participant study identifier and no personal data. All data will be returned to Oxford for statistical analyses.

The study team intend to link data from participant DNA analyses to NHS Digital data. As demonstrated by the HPS long-term follow-up results, linkage to longer term outcomes (including those identified from NHS Digital data) is of considerable importance in fully understanding response to statins and cardiovascular risk and its consequences as a whole. Furthermore, genetic data may be able to add insights into our understanding of the effects of statins, determinants and mechanisms of risk factors and disease outcomes, that can contribute to the development of better patient treatments and precision medicine approaches.

The processing of the data under this Agreement will be carried out in accordance with Article 6(1)(e) as processing is necessary for a task carried out in the public interest. The aim of the study is to provide reliable evidence about the very long term effects of cholesterol-lowering treatments in the UK population and is therefore in the public interest.

Processing of the data is in accordance with Article 9(2)(j) exemption, i.e. that the processing of the data is necessary for scientific research. The scientific research aim of the study is to provide reliable evidence about the very long term effects of cholesterol-lowering treatments on important health outcomes.

The research team have taken the opinion of the South Central – Oxford B Research Ethics Committee, who have granted approval to the study in its current form. Furthermore, potential harm to the public will be limited by data pseudonymisation, data minimisation, and technical and organisational safeguards. Analysis will occur in a safe NHS Data Security and Protection (DSP) Toolkit compliant environment and identifiers will not be included in the analysis. Although the study is not still in direct contact with participants, at the time the study team asked for the opinion of available participants in these and similar studies, they agreed that the use of the data in the way that the study team have proposed is reasonable.

All processing of NHS Digital data will take place within the University of Oxford, therefore Oxford are listed as sole data processor under this Agreement.

One of the Chief Investigators is employed by both Gloucestershire Hospitals NHS Foundation Trust and the University of Oxford. This Chief Investigator will analyse the data under this Agreement solely in their role as a substantive employee of the University of Oxford, and therefore Cheltenham General Hospital is not listed as a data controller under this Agreement.

Funding for this study is provided internally from the University of Oxford.

The primary outputs from this data will be academic, and will include submissions to peer reviewed journals such as New England Journal of Medicine and the Lancet, and conferences such as the American Heart Association meeting.

The study team will only present data at an aggregated level with small numbers suppressed. The study team will present actual and modelled data in graphical and tabular format.

The NDPH contributes widely to national and international health policy, particularly in the area of vascular risk prevention. It contributes to debate with academic papers, conference participation, lectures to the public and advice to government (including NHS Digital). Examples of the impact of the work performed by NDPH up to 2014 is available from: https://results.ref.ac.uk/(S(ep5gbndxsprqnc0kork3mjyu))/Submissions/Impact/728

The study team will share outputs via the following channels:

- Study website

- Open lectures and talks

- Exhibition at public events

- Posters

- Press/media engagement and other public promotion of the research

The data controller aims to issue the next publication by the end of 2023.

All outputs will only contain results in highly aggregated format and as statistical summaries and measures of association. Small numbers will be suppressed in line with the HES Analysis Guide. Record level information will not be released to any third party.

Participants have been provided with updates throughout the years since the study began in 1994. Newsletters were mailed to participants between 1996 and 2000. In 2001, a letter was mailed which included information about the main trial results and details of the HPS study website. This website has been regularly updated since 2001 and in 2020 the study team uploaded two new documents to the study, one was a lay summary for the HPS follow-up study and one was a privacy notice for the follow-up study.

Outputs for the HPS study so far include multiple publications in peer-reviewed journals and presentations at international conferences. Information about the publications can be found on the HPS website https://www.ctsu.ox.ac.uk/research/hps

This study has strongly influenced the labelling of statin medication internationally, treatment guidelines, and the resulting changes in prescribing have contributed to reductions in mortality and morbidity from heart attack and blocked arteries (‘ischaemic stroke’) in many countries.

The main results of HPS (published in 2002) clearly showed that allocation to cholesterol-lowering therapy with simvastatin was associated with a 20% reduction in the risk of heart attacks and stroke, and a reduction in all-cause mortality, chiefly driven by a very significant reduction in vascular deaths. These results changed clinical practice worldwide, and statins are now prescribed widely to many millions of patients. The following outcomes for the study influenced NICE Guidelines within the UK:

1) Statin therapy significantly reduces the incidence of coronary and other major vascular events to a similar extent, irrespective of KIF6 genotype. Consequently, the use of KIF6 genotyping to guide statin therapy is not warranted.

2) These findings provide considerable reassurance that lowering total cholesterol concentrations by more than 1 mmol/L for an average of 5 years does not produce adverse effects on non-vascular mortality or cancer incidence. Moreover, among the many different types of high-risk individual studied, simvastatin 40 mg daily consistently produced substantial reductions in vascular (and, hence, all-cause) mortality, as well as in the rates of non-fatal heart attacks, strokes and revascularisation procedures.

3) Treatment with statins is cost effective in a wider population than is routinely treated at present.

4) Adding simvastatin to existing treatments safely produces substantial additional benefits for a wide range of high-risk patients, irrespective of their initial cholesterol concentrations. Allocation to 40 mg simvastatin daily reduced the rates of myocardial infarction, of stroke, and of revascularisation by about one-quarter. After making allowance for non-compliance, actual use of this regimen would probably reduce these rates by about one-third. Hence, among the many types of high-risk individual studied, 5 years of simvastatin would prevent about 70-100 people per 1000 from suffering at least one of these major vascular events (and longer treatment should produce further benefit). The size of the 5-year benefit depends chiefly on such individuals' overall risk of major vascular events, rather than on their blood lipid concentrations alone.

5) The present study provides direct evidence that cholesterol-lowering therapy is beneficial for people with diabetes even if they do not already have manifest coronary disease or high cholesterol concentrations. Allocation to 40 mg simvastatin daily reduced the rate of first major vascular events by about a quarter in a wide range of diabetic patients studied. After making allowance for non-compliance, actual use of this statin regimen would probably reduce these rates by about a third. For example, among the type of diabetic patient studied without occlusive arterial disease, 5 years of treatment would be expected to prevent about 45 people per 1000 from having at least one major vascular event (and, among these 45 people, to prevent about 70 first or subsequent events during this treatment period). Statin therapy should now be considered routinely for all diabetic patients at sufficiently high risk of major vascular events, irrespective of their initial cholesterol concentrations.

6) Much larger numbers of people in the present study suffered a stroke than in any previous cholesterol-lowering trial. The results demonstrate that statin therapy rapidly reduces the incidence not only of coronary events but also of ischaemic strokes, with no apparent effect on cerebral haemorrhage, even among individuals who do not have high cholesterol concentrations. Allocation to 40 mg simvastatin daily reduced the rate of ischaemic strokes by about one-quarter and so, after making allowance for non-compliance in the trial, actual use of this regimen would probably reduce the stroke rate by about a third. HPS also provides definitive evidence that statin therapy is beneficial for people with pre-existing cerebrovascular disease, even if they do not already have manifest coronary disease.

Mortality, cancer registration, NHS registration and Hospital Episode Statistics data were supplied to the University of Oxford for the purpose of a research study called 'The Heart Protection Study (HPS)'.

This Data Sharing Agreement permits the University of Oxford to retain and process the data for the duration of the Agreement. Permission to retain the data for an interim period is a practical step to enable the data recipient to comply with the necessary requirements in order to secure a longer-term Data Sharing Agreement permitting further retention and reuse of the data for approved purposes.

The following information provides background information on the purpose of the original study:

The Heart Protection Study (HPS) began in 1994. This randomized trial demonstrated that lowering Low Density Lipoprotein (LDL) cholesterol with statins reduces vascular morbidity and mortality, and, as a result of the results published in 2002, such medications are now widely prescribed.

The MRC/BHF Heart Protection Study (HPS) has demonstrated that cholesterol-lowering with 40mg simvastatin daily for an average of 5 years substantially reduces the risk of major vascular events in a much wider range of individuals than previously thought to benefit from such treatment. The reductions in risk increased throughout the scheduled treatment period, suggesting that more prolonged treatment might result in larger benefits. No adverse effects on non-vascular mortality or cancer were observed during the statin treatment period. HPS also assessed the effects of antioxidant vitamin supplementation (vitamins E, C and beta-carotene), with no significant benefits or hazards observed during the 5-year treatment period. Prolonged follow-up of participants via routinely collected health records and central registries allows reliable assessment of any delayed effects on major outcomes.

Planned comparisons of outcome during extended follow-up:

The primary objectives are to assess the effects of randomly allocating 20,536 high-risk patients to receive an average of 5 years of 40mg simvastatin daily versus placebo (and, in a “2x2 factorial design”, of antioxidant vitamin supplementation versus placebo) on major vascular events (MVEs) during 16 years of further follow-up after the end of the scheduled study treatment period (2001). Secondary objectives are to assess the effects during this period of extended follow-up: (i) on major coronary events, on non-fatal and fatal strokes, and on coronary and non-coronary revascularisations; (ii) on deaths from all vascular and from all non-vascular causes; (iii) on cancers at all sites (excluding non-melanoma skin cancer); and (iv) on MVEs during each separate year of follow-up and in various subcategories of patients (as pre-specified in the Data Analysis Plan for the scheduled treatment period: see www.hpsinfo.org). Tertiary objectives will include assessment during extended follow-up of the effects on other major adverse outcomes (e.g. site-specific cancer; cerebral haemorrhage, hospitalisation for angina and for various other causes).

All comparisons will involve logrank analyses of the first occurrence of particular events during the extended follow-up period among all those allocated simvastatin versus all those allocated matching placebo and, separately, among all those allocated vitamins versus all those allocated matching placebo (i.e. they will be “intention-to-treat” analyses). Tests for heterogeneity will be used to assess whether the proportional effects observed in specific subcategories differ clearly from the overall effects.

One aim of the post-trial follow-up of the study is to add to the limited available evidence about the long-term safety of statins and antioxidant vitamins. Deleterious effects (e.g. cancer incidence) may take many years to emerge.

The post-trial follow-up of all surviving HPS participants for cause specific mortality and cancer incidence allows direct assessment of the long-term safety and efficacy of lowering LDL cholesterol by around 1.0 mmol/L for 5 years with 40 mg simvastatin.

In addition, whilst allocation to antioxidant vitamin supplements had no effect on vascular disease, non-vascular mortality or cancer incidence during the scheduled study period, it is possible that benefit (or harm) associated with antioxidant vitamin allocation may emerge with more prolonged follow-up.

The expected outputs from this data will be academic, and will include submissions to peer reviewed journals such as New England Journal of Medicine and the Lancet, and conferences such as the American Heart Association meeting.

The study team will only present data at an aggregated level with small numbers suppressed. The study team will present actual and modelled data in graphical and tabular format.

All outputs will only contain results in highly aggregated format and as statistical summaries and measures of association. Small numbers will be suppressed in line with the HES Analysis Guide. Record level information will not be released to any third party.

The HPS demonstrated that lowering Low Density Lipoprotein (LDL) cholesterol with statins reduces vascular morbidity and mortality, and, as a result of the results published in 2002, such medications are now widely prescribed.

For various reasons it has not been possible to analyse HES data supplied previously. Initially, the data were more complex than had been anticipated (particularly given the relatively morbid population included in HPS) and it was difficult to decipher incident (i.e. new) events from prevalent disease. Then, problems with data flow from death registries meant it was not possible to censor the study population.

Both these problems have now been overcome, but analysis of the data we currently hold without updating with more recent years would invite speculation from reviewers and readers as to why we had not included all available data.

Mortality, cancer registration, NHS registration and Hospital Episode Statistics data were supplied to the University of Oxford for the purpose of a research study called 'The Heart Protection Study (HPS)'.

This Data Sharing Agreement permits the University of Oxford to retain and process the data for the duration of the Agreement. Permission to retain the data for an interim period is a practical step to enable the data recipient to comply with the necessary requirements in order to secure a longer-term Data Sharing Agreement permitting further retention and reuse of the data for approved purposes.

The following information provides background information on the purpose of the original study:

The Heart Protection Study (HPS) began in 1994. This randomized trial demonstrated that lowering Low Density Lipoprotein (LDL) cholesterol with statins reduces vascular morbidity and mortality, and, as a result of the results published in 2002, such medications are now widely prescribed.

The MRC/BHF Heart Protection Study (HPS) has demonstrated that cholesterol-lowering with 40mg simvastatin daily for an average of 5 years substantially reduces the risk of major vascular events in a much wider range of individuals than previously thought to benefit from such treatment. The reductions in risk increased throughout the scheduled treatment period, suggesting that more prolonged treatment might result in larger benefits. No adverse effects on non-vascular mortality or cancer were observed during the statin treatment period. HPS also assessed the effects of antioxidant vitamin supplementation (vitamins E, C and beta-carotene), with no significant benefits or hazards observed during the 5-year treatment period. Prolonged follow-up of participants via routinely collected health records and central registries allows reliable assessment of any delayed effects on major outcomes.

Planned comparisons of outcome during extended follow-up:

The primary objectives are to assess the effects of randomly allocating 20,536 high-risk patients to receive an average of 5 years of 40mg simvastatin daily versus placebo (and, in a “2x2 factorial design”, of antioxidant vitamin supplementation versus placebo) on major vascular events (MVEs) during 16 years of further follow-up after the end of the scheduled study treatment period (2001). Secondary objectives are to assess the effects during this period of extended follow-up: (i) on major coronary events, on non-fatal and fatal strokes, and on coronary and non-coronary revascularisations; (ii) on deaths from all vascular and from all non-vascular causes; (iii) on cancers at all sites (excluding non-melanoma skin cancer); and (iv) on MVEs during each separate year of follow-up and in various subcategories of patients (as pre-specified in the Data Analysis Plan for the scheduled treatment period: see www.hpsinfo.org). Tertiary objectives will include assessment during extended follow-up of the effects on other major adverse outcomes (e.g. site-specific cancer; cerebral haemorrhage, hospitalisation for angina and for various other causes).

All comparisons will involve logrank analyses of the first occurrence of particular events during the extended follow-up period among all those allocated simvastatin versus all those allocated matching placebo and, separately, among all those allocated vitamins versus all those allocated matching placebo (i.e. they will be “intention-to-treat” analyses). Tests for heterogeneity will be used to assess whether the proportional effects observed in specific subcategories differ clearly from the overall effects.

One aim of the post-trial follow-up of the study is to add to the limited available evidence about the long-term safety of statins and antioxidant vitamins. Deleterious effects (e.g. cancer incidence) may take many years to emerge.

The post-trial follow-up of all surviving HPS participants for cause specific mortality and cancer incidence allows direct assessment of the long-term safety and efficacy of lowering LDL cholesterol by around 1.0 mmol/L for 5 years with 40 mg simvastatin.

In addition, whilst allocation to antioxidant vitamin supplements had no effect on vascular disease, non-vascular mortality or cancer incidence during the scheduled study period, it is possible that benefit (or harm) associated with antioxidant vitamin allocation may emerge with more prolonged follow-up.

The expected outputs from this data will be academic, and will include submissions to peer reviewed journals such as New England Journal of Medicine and the Lancet, and conferences such as the American Heart Association meeting.

The study team will only present data at an aggregated level with small numbers suppressed. The study team will present actual and modelled data in graphical and tabular format.

All outputs will only contain results in highly aggregated format and as statistical summaries and measures of association. Small numbers will be suppressed in line with the HES Analysis Guide. Record level information will not be released to any third party.

The HPS demonstrated that lowering Low Density Lipoprotein (LDL) cholesterol with statins reduces vascular morbidity and mortality, and, as a result of the results published in 2002, such medications are now widely prescribed.

For various reasons it has not been possible to analyse HES data supplied previously. Initially, the data were more complex than had been anticipated (particularly given the relatively morbid population included in HPS) and it was difficult to decipher incident (i.e. new) events from prevalent disease. Then, problems with data flow from death registries meant it was not possible to censor the study population.

Both these problems have now been overcome, but analysis of the data we currently hold without updating with more recent years would invite speculation from reviewers and readers as to why we had not included all available data.

Mortality, cancer registration, NHS registration and Hospital Episode Statistics data were supplied to the University of Oxford for the purpose of a research study called 'The Heart Protection Study (HPS)'.

This Data Sharing Agreement permits the University of Oxford to retain and process the data for the duration of the Agreement. Permission to retain the data for an interim period is a practical step to enable the data recipient to comply with the necessary requirements in order to secure a longer-term Data Sharing Agreement permitting further retention and reuse of the data for approved purposes.

The following information provides background information on the purpose of the original study:

The Heart Protection Study (HPS) began in 1994. This randomized trial demonstrated that lowering Low Density Lipoprotein (LDL) cholesterol with statins reduces vascular morbidity and mortality, and, as a result of the results published in 2002, such medications are now widely prescribed.

The MRC/BHF Heart Protection Study (HPS) has demonstrated that cholesterol-lowering with 40mg simvastatin daily for an average of 5 years substantially reduces the risk of major vascular events in a much wider range of individuals than previously thought to benefit from such treatment. The reductions in risk increased throughout the scheduled treatment period, suggesting that more prolonged treatment might result in larger benefits. No adverse effects on non-vascular mortality or cancer were observed during the statin treatment period. HPS also assessed the effects of antioxidant vitamin supplementation (vitamins E, C and beta-carotene), with no significant benefits or hazards observed during the 5-year treatment period. Prolonged follow-up of participants via routinely collected health records and central registries allows reliable assessment of any delayed effects on major outcomes.

Planned comparisons of outcome during extended follow-up:

The primary objectives are to assess the effects of randomly allocating 20,536 high-risk patients to receive an average of 5 years of 40mg simvastatin daily versus placebo (and, in a “2x2 factorial design”, of antioxidant vitamin supplementation versus placebo) on major vascular events (MVEs) during 16 years of further follow-up after the end of the scheduled study treatment period (2001). Secondary objectives are to assess the effects during this period of extended follow-up: (i) on major coronary events, on non-fatal and fatal strokes, and on coronary and non-coronary revascularisations; (ii) on deaths from all vascular and from all non-vascular causes; (iii) on cancers at all sites (excluding non-melanoma skin cancer); and (iv) on MVEs during each separate year of follow-up and in various subcategories of patients (as pre-specified in the Data Analysis Plan for the scheduled treatment period: see www.hpsinfo.org). Tertiary objectives will include assessment during extended follow-up of the effects on other major adverse outcomes (e.g. site-specific cancer; cerebral haemorrhage, hospitalisation for angina and for various other causes).

All comparisons will involve logrank analyses of the first occurrence of particular events during the extended follow-up period among all those allocated simvastatin versus all those allocated matching placebo and, separately, among all those allocated vitamins versus all those allocated matching placebo (i.e. they will be “intention-to-treat” analyses). Tests for heterogeneity will be used to assess whether the proportional effects observed in specific subcategories differ clearly from the overall effects.

One aim of the post-trial follow-up of the study is to add to the limited available evidence about the long-term safety of statins and antioxidant vitamins. Deleterious effects (e.g. cancer incidence) may take many years to emerge.

The post-trial follow-up of all surviving HPS participants for cause specific mortality and cancer incidence allows direct assessment of the long-term safety and efficacy of lowering LDL cholesterol by around 1.0 mmol/L for 5 years with 40 mg simvastatin.

In addition, whilst allocation to antioxidant vitamin supplements had no effect on vascular disease, non-vascular mortality or cancer incidence during the scheduled study period, it is possible that benefit (or harm) associated with antioxidant vitamin allocation may emerge with more prolonged follow-up.

Several major publications describing the long-term safety and efficacy of 5 years lipid-lowering therapy with simvastatin and, separately, antioxidant vitamins are planned. Details will be provided in a future application to NHS Digital.

All outputs will be aggregated with small numbers suppressed in line with the HES Analysis Guide.

The HPS demonstrated that lowering Low Density Lipoprotein (LDL) cholesterol with statins reduces vascular morbidity and mortality, and, as a result of the results published in 2002, such medications are now widely prescribed.

For various reasons it has not been possible to analyse HES data supplied previously. Initially, the data were more complex than had been anticipated (particularly given the relatively morbid population included in HPS) and it was difficult to decipher incident (i.e. new) events from prevalent disease. Then, problems with data flow from death registries meant it was not possible to censor the study population.

Both these problems have now been overcome, but analysis of the data we currently hold without updating with more recent years would invite speculation from reviewers and readers as to why we had not included all available data.