NPC Archive Item: QRISK2 – a valid tool for predicting cardiovascular risk?

NOTE – This is an archive post from the NPC and has not been updated since first publication. Therefore, some hyperlinks may no longer be working.
MeReC Rapid Review NPC Logo

Hippisley-Cox J, Coupland C, Vinogradova Y, et al. Predicting cardiovascular risk in England and Wales: prospective derivation and validation of QRISK2. BMJ 2008;336;1475–82

QRISK2 may provide a more accurate assessment of cardiovascular risk than the modified Framingham scoring system for people in England and Wales, particularly for those above the 20% 10-year cardiovascular risk threshold.

Current National Institute of Health and Clinical Effectiveness (NICE) guidance should still be followed. This advises that people aged 40-75 years who are at high risk of cardiovascular disease (CVD) should have their 10-year risk assessed using the Framingham 1991 scoring system as advised by the current NICE guideline. If NICE should update their guidance then we would incorporate this into future NPC and MeReC publications.

What is the background to this?
Unlike CVD risk assessment tools based on Framingham, the QRISK2 tool is based on English and Welsh data. It also incorporates ethnicity, social deprivation and additional relevant clinical conditions aiming to provide a more accurate assessment of risk. The NICE lipid modification guideline was published in May 2008 prior to publication of the QRISK2 data. The guidelines emphasise the limitation of Framingham risk scores in their exclusion of risk factors, such as ethnic origin and socioeconomic status, and how current adaptations to take account of these may result in incorrect risk assessment. Recommendations were made by the guideline development group for further research and reconsideration of alternative risk assessment tools. QRISK2 is a revised version and in this paper is reported as giving a more accurate assessment of people at high risk of CVD.

What does this study claim?
This study reports QRISK2 to be more accurate in identifying those who are at high risk of CVD than Framingham in England and Wales. Validation showed QRISK2 had improved discrimination and calibration compared with the Framingham equation. Calibration relates to how close the predicted risk is to the observed risk. More importantly, discrimination is the ability of the tool to differentiate between people who will have an event and those people who will not over a defined period of time (often five to ten years). Those identified by QRISK2 at the >20% threshold were at a higher risk of a cardiovascular event compared to those identified with a Framingham score. Overall at the >20% 10-year CVD risk threshold, QRISK2 would predict 10.4% of patient to be at high risk compared to 14.9% with Framingham. Of the patients classified as high risk with QRISK2 15.3% would be reclassified as low risk with Framingham scores. Of those classified as high risk with Framingham 41% would be classified as low risk with QRISK2.

How does this relate to other studies?
Current NICE guidance states the 10-year CVD risk should be formally assessed by a Framingham-based scoring system to take into account age, sex, systolic blood pressure, total cholesterol, HDL cholesterol, smoking status and the presence of left ventricular hypertrophy. NICE suggests that the formally calculated CVD risk should be multiplied by 1.4 for South Asian men and by 1.5 in people with a first-degree relative with a history of premature* coronary heart disease (CHD).  The formally calculated CVD risk should be multiplied by a factor between 1.5 and 2 if more than one first-degree relative has a history of premature CHD. Evidence does not support an adjustment for women of south Asian origin. Many other factors that affect assessment of CVD risk are not included in the modified Framingham equation, including socioeconomic status, obesity, recent smoking cessation and existing antihypertensive/lipid-lowering treatments. The guideline development group emphasised that risk estimation tools only provide an approximation of CVD risk and should be used together with clinical judgment to guide management.

* Premature is defined as age at onset younger than 55 years in fathers, sons or brothers or younger than 65 years in mothers, daughters or sisters.

So what?
In their guidance NICE highlights the limitations of the Framingham equation and called for further research into risk assessment tools (QRISK and ASSIGN) in an English and Welsh population. This study goes some way to fulfilling these objectives. The study indicates that ethnicity, socioeconomic status and additional clinical conditions affect CVD risk assessment and that any risk assessment tool should take account of these factors to provide a personalised assessment of risk. QRISK2 provides a tool for assessing risk that takes account of these factors. The authors acknowledge that QRISK2 may hold a “home advantage” as validation was conducted with English and Welsh data. This will affect generalisability to other populations.

Clinicians should continue to follow current NICE guidelines. Should NICE guidance change as a result of this study it will be incorporated in future NPC publications.

Study details
Design: Prospective open cohort study with data from 531 general practices in England and Wales contributing to the national QRESEARCH database.

Patients: 2.3 million patient aged 34–74 years with 2.22 million who were white or whose ethnic group was not recorded, 22,013 south Asian, 11,595 black African, 10,402 black Caribbean and 19,792 from Chinese or other Asian or other ethnic groups (derivation and validation cohort).

Intervention : QRISK2 is a modified QRISK1 algorithm for CVD risk assessment to adjust for ethnicity, social deprivation, type 2 diabetes, treated hypertension, rheumatoid arthritis, renal disease and atrial fibrillation.

Comparison: The predicted and observed risks of CVD over 10-years for QRISK2 were compared with those of QRISK1 and the modified Framingham 1991 equation (as recommended by NICE).

Outcomes: First (incident) diagnosis of CVD (coronary heart disease, stroke and transient ischemic attack). Risk factors included self assigned ethnicity, age, sex, smoking status, systolic blood pressure, ratio of total serum cholesterol: high density lipoprotein cholesterol, body mass index, family history of coronary heart disease in first degree relative under 60 years, Townsend deprivation score, treated hypertension, type 2 diabetes, renal disease, atrial fibrillation and rheumatoid arthritis.

Results: Calibration and discrimination statistics indicated that QRISK2 was marginally superior to QRISK1 and both models were superior to the modified Framingham scoring method. For example QRISK2 explained 43% of the variation between predicted and observed risk in women and 38% in men compared to 39% and 35% with the Framingham score (R2 statistic). Overall at the >20% cardiovascular risk threshold QRISK2 would predict 10.4% of patient to be at high risk compared to 14.9% with Framingham. Of the patients at the >20% threshold with QRISK2 15.3% would be reclassified as low risk with Framingham scores. Those identified by QRISK2 at the >20% threshold were at a higher risk of a cardiovascular event compared to those identified with a Framingham score. The annual incident rate at the >20% threshold with QRISK2 was 30.6 per 1000 person years (95% Confidence Interval (CI) 29.8 to 31.5) for women and 32.5 per 1000 person years (95% CI 31.9 to 33.1) for men. With Framingham the annual incident rate for woman was 25.7 per 1000 person years (95% CI 25.0 to 26.3) and 26.4 (26.0 to 26.8) for men.

Sponsorship: No external funding. The authors were funded as part of their clinical or academic positions and meeting expenses were met by the University of Nottingham.

Please comment on this blog in the NPC discussion rooms, or using our feedback form

The primary author of this blog is Elisabeth Maskrey, 3rd year medical student at Hull York Medical School.