Primary Intentions

Introduction

This changed when I was undertaking my training in 2023 where new recommendations were made where the Framingham equation was replaced with the newer AusCVDRisk Calculator for primary prevention.

The new calculator can be found at https://www.cvdcheck.org.au/

There is some nuance to risk assessment. The pathway can be very algorithmic in nature.

What I want to examine is the evidence behind it, what are the advantages and disadvantages we have, what do you do for the grey area and where to from here?

Note: if this is too much waffle, I have put a summary at the end

The Framingham Risk score – old and discarded

This was the tool used previously in Australia. The origin of it is a epidemiological cohort study that continues on from 1948 in Framingham, Masschusetts, USA. It continues today with its most recent tool for CVD risk stratification; the 2018 Prevention Guidelines Tool CV Risk Calculator. The Americans use it. We don’t anymore.

This tool enables us to estimate the 10-risk and lifetime risks of cardiovascular disease (specifically atherosclerotic of course).

The disadvantage of this tool is as follows:

1. USA data (we’re Australians and we have drop bears
2. Primarily used for coronary heart disease (mainly myocardial infarction) where peripheral vascular disease and stroke were not taken into account as much
3. it does not take into account chronic kidney disease markers
4. Does not take into account demographic markers
5. Only takes 10 years into account.

You can find the calculator here: https://www.framinghamheartstudy.org/fhs-risk-functions/cardiovascular-disease-10-year-risk/

What is interesting is they also have a 30 year calculator as well which I do not think we have much teaching on nor do we think much about this. This calculator uses a study that followed 4506 participants (which half were women) that did not have cardiovascular disease from 1971 to 1974 and examined the variables involved. The predictors of heart disease remain the same and overall I don’t think I would be advocating for its use. I think it’s a niche tool that can be used if a patient is really motivated to make changes and see their data: for clinical decision making I wouldn’t rely on it at all.

AusCVDRisk Calculator – behind the scenes

This tool was developed for the Australian population where the National Heart Foundation of Australia was contracted by the Department of Health and Aged Care.

If you get the appendix and the individuals involved boy oh boy it’s robust. You have leading experts in multiple fields including epidemiology, endocrinology, cardiology and more.

They are also very transparent in how decisions were made based on evidence-based medicine and they even grade the certainty of evidence.

They even leaned on our neighbouring New Zealander’s cohort study on CVD through the utilization of their NZ PREDICT-1 equation (if you’re interested have a look here https://www.vareanz.auckland.ac.nz/predict1-riskscores/)

Overall, it’s quite solid in what it delivers for primary prevention.

The key issue I have with it it’s a 5 year predictor. I imagine over time the data generated from the Heart Foundation Australia will give us a tool to stretch it more into the future. To address this weakness it asks us the GP to reassess risk at 5 year intervals for low risk, and lower for the intermediate and high risk patients. There’s a few more issues which I will go into more in depth later.

AusCVDRisk Calculator – the meaty bits

When confronted to the calculator to use for a patient it’s taking into account multiple variables that we know are linked with CVD. It also takes into account the postcode – ever won a postcode lottery? Live in Toorak. Socioeconomic disadvantage is taken into account. Key risk factors are noted.

The disadvantages are laid out as well with reclassification factors – family history, certain ethnicities, and highlighting that patients with certain other health issues.

Now, to summarize briefly their recommendations: low risk = lifestyle modifications only, medium risk = lifestyle +/- meds, high risk = definitely recommending medications.

There’s people with cholesterol >7.5 total and LDL > 5.0 or CKD Stage 3b or lower who automatically get high risk as well. There’s studies in the past for the cholesterol parameters that I dug up and they found these points were cut offs where the CVD risk became high (3-4x higher): this included the Framingham Heart Study, the INTERHEART study and the Copenhagen City Heart Study. Don’t forget the familial hypercholesterolaemia cohort as well.

For chronic kidney disease stage 3b or lower, four big studies (Go et al, NEJM 2004;351:1296-1305), (Chronic Kidney Disease Prognosis Consortium, Lancet), (Fox et al, NEJM 2012;367:240-249) and (Matsushita et al, Circulation 2020;141:1242-1254) all demonstrate increasing CVD risk with declining renal function. The key point of definitely intervening is stage 3b, but you can argue 3a as well (which the ACVDR takes into account).

This is the first point of nuance to consider.

1. Ethnicity – Maori, Pacific Islander or South Asian ethnicity
2. Family history of premature CVD – males < 55, females < 65 (first degree)
3. Chronic kidney disease – getting closer to 3a increases risk. it also goes into A2 or A3 as markers of higher risk as well.
4. People living with severe mental illness – I think we also under-appreciate this and often get lost in the forest of managing their mental health.

This area I find tricky if the patient overall is low risk with these factors. What do I do then? Do I reclassify them to intermediate risk? Is this intermediate risk equivalent to a person who doesn’t have these factors but simply due to their age and normotension at 130 with a ratio of cholesterol of 5? Should I start the person on medications?

My default position becomes could this patient then benefit from a CT coronary calcium score for more risk stratification if they fall into the intermediate zone or low risk with concerning features such as family history. Which then follows a similar line of thinking with what is their calcium score and then low risk, intermediate risk and high risk. I’ll write more on this topic in another article.

So the answer is clear if the patient is low risk and no other factors or high risk which means straight onto medications. I won’t dive into the high risk area as much in my discussion.

AusCVDRisk Calculator – the intermediate risk zone

So why is it the recommendations say consider this area a discussion with the patient?

Table 1: Primary Prevention Trials in Intermediate Risk

Study	Population	Baseline 5-Year Risk	Key Findings	Reference
JUPITER (2008)	17,802 adults with LDL <3.4 mmol/L and hs-CRP ≥2 mg/L	Mean ≈ 7%	Rosuvastatin 20 mg daily reduced MI, stroke, and CV death by 44% RRR; ARR ≈ 1.2%, NNT = 83/5 years	Ridker PM et al. NEJM 2008;359:2195-2207
HOPE-3 (2016)	12,705 men ≥55 y / women ≥65 y with ≥1 risk factor	Median 5-year risk ≈ 8%	Rosuvastatin 10 mg ↓ CV events by 24% RRR; ARR ≈ 1.1%, NNT = 91/5 years	Yusuf S et al. NEJM 2016;374:2021-2031
ASCOT-LLA (2003)	Hypertensive, non-diabetic adults, mean cholesterol 6 mmol/L	~10%	Atorvastatin 10 mg ↓ non-fatal MI/fatal CHD by 36% RRR; ARR ≈ 1.1%, NNT ≈ 91/5 years	Sever PS et al. Lancet 2003;361:1149-1158
WOSCOPS (1995)	Men with moderate hyperlipidaemia, no CVD	≈ 9%	Pravastatin ↓ CHD by 31% RRR; ARR ≈ 2.4%, NNT = 42/5 years	Shepherd J et al. NEJM 1995;333:1301-1307

Table 2: Meta-analyses Supporting Benefit Across Risk Bands

Analysis	Risk Groups Studied	Result Summary	Reference
CTT Collaboration 2012	27 RCTs, 134 000 participants	Every 1 mmol/L LDL reduction = 25% RRR in major vascular events across all baseline risks. For 5-year risk < 10%, ARR ≈ 1–2% (NNT ≈ 50–100)	Lancet 2012;380:581-590
Mihaylova et al. 2012	Pooled > 175 000 participants	In primary prevention, for each 1 mmol/L LDL ↓, 22% fewer major vascular events; mortality reduction appears from baseline 10-year risk > 7.5%	Lancet 2012;380:581-590
Naci et al. BMJ 2013	29 trials, 80 711 pts	NNT ≈ 49 for all-cause mortality over 5 years in moderate-risk primary prevention	BMJ 2013;347:f6123
Silverman et al. JAMA 2016	49 RCTs, 312 000 pts	Statin use in intermediate risk lowers CV events (RR 0.73) and all-cause mortality (RR 0.91)	JAMA 2016;316(19):1975-1989

There’s probably more studies out there but overall, what we can see is statins, if introduced, has a relative risk reduction of about 25% irrespective of the risk.

So why not automatically put it for everyone in intermediate risk? Why only reserved for higher risk individuals?

Another way to think about it is what is the number needed to treat to stop 1 cardiovascular event over 5 years? For intermediate risk it’s around 42-91 depending on which study you use above.

I have to treat close to 100 patients to stop 2 cardiovacular events at best and at worst 1 cardiovascular event. We can see for high risk people the number of events we prevent becomes higher per 100 people treated and thus the default position is to treat. So this is a modest reduction

Tricky isn’t it? I mean we can argue stopping 2 heart attacks out of 100 people treated on statin is beneficial of course, but we also need to think about the number needed to harm with statin therapy. The harms of statin therapy if we look at the numbers is muscle associated side effects (around 1 in 20). What is important to note as well is a randomized large-scale clinical trial demonstrated that statins have a dose dependent effect on worsening hyperglycaemia in patients (note it is SMALL, and there is still a reduction in cardiovascular risk overall with this tiny increase risk); effects of statin therapy on diagnoses of new-onset diabetes and worsening glycaemia in large-scale randomised blinded statin trials: an individual participant data meta-analysis

So if patients are in the intermediate zone I think I would be inclined to say there’s a 25% reduction in cardiovascular events and based on your risk a NNT is somewhere between 42-91, probably a much narrower range if we dig up more studies, but it’s not mandatory. The ACVDR evidence digs into this a lot as well. Of course, it’s the patient’s decision but we can use this to help them make decisions accordingly to reduce their risk of heart disease. Would I put these patients on statins by default? I let the patient make the decision and I steer them to make lifestyle changes because that will improve not only their cardiovascular disease risk but their overall wellbeing and health.

Aspirin follows a similar line of thinking with NNT and NNH – unfortunately the numbers are so close together can we really argue it’s beneficial? Even at high risk individuals it’s still a tricky line to consider.

This was a bit of a rant but I needed to look at the data myself to convince myself the ACVDR consensus that it’s a discussion rather than push medications immediately is reasonable. Even looking at the intermediate risk data, you get insight into why it is then that you treat for high risk individuals and not treat for low risk individuals. It’s all down to the NNT. But don’t forget NNH

What about low CVD risk? Should we treat with a statin?

I think we have been a bit trigger happy with statin prescriptions to stop CVD – it is up there in mortality. It comes down to the practitioner who prescribed them facing the number and going any LDL > 2.0 is unacceptable and over-estimating the benefit statins have to prevent harm from CVD.

If the LDL > 5.0 or total chol is 7.5 then the evidence indicates the 2.5-3x risk of CVD makes it unacceptable to not offer statin therapy for these individuals – don’t forget the familial hypercholesterolaemia patients either who fall into this.

This is an opportunity to try and identify the patient’s motivational level of change and try to get them to make goals to change their lifestyle factors.

I found the resource theNNT.com quite helpful to look at. For persons at low risk of cardiovascular disease, with the references and resources they used, they have demonstrated there is no statistically significant mortality benefit if we treat with statins. You need to treat about 200 people to avoid 1 non-fatal myocardial infarction and about 300 people to avoid 1 non-fatal stroke. Talk about inefficient. Don’t forget NNH as well with the slight increased amount of hyperglycaemia and muscle related side effects we discussed earlier.

What other tools do we have?

Apolipoprotein B, lipoprotein(A), polygenic risk scores (so genetic variant testing) and inflammatory markers (hs-CRP in particular potentially with others showing some promise such as GDF-15, suPAR and IL-6) will have roles in the future.

We already see Canadian guidelines emerging for ApoB and other countries as well. I will most likely look into this as I imagine in a few years, up to 5, this will soon be incorporated into Australian General Practice. For now, I know of colleagues who are already diving into testing these factors and managing patients using them.

Conclusions

1. The new way forward for Australian patients for primary prevention of CVD is to use the AusCVDRisk Calculator
2. Key risk factors to not forget:
   • increasing CVD risk with declining renal function – progressive stages and progressive albuminuria increases risk, with stage 3b or lower making it automatically high risk
   • Familial hypercholesterolaemia – don’t forget this
   • First Nations people – unfortunately disproportionately higher rates of mortality and morbidity from multitudes of conditions
   • Maori, Pacific Islander or South Asian ethnicity
   • People living with severe mental illness – you have seen these patients on multiple psychotropes, don’t get lost in the forest of managing their mental health and take a step back to see the whole area
   • Family history of CVD
3. Intermediate risk individuals
   • Challenging situation
   • Being aware of the above variables will guide decision alongside the patient’s BP and chol profile
   • The reduction in CVD is modest here (NNT from above around 41-92) looking at the studies but if multiple risk factors, not accounted for by the AusCVDRisk calculation, are present then I would be more inclined to treat especially if the risk is closer to high risk than low
   • The CT calcium score can be used here – I’ll write more about this separately
4. Don’t forget lifestyle modification – assessing motivational stage of change and guide the patient to making their own goals to meet to improve their health
5. Low risk CVD patients, without severe hypertension and TC < 7.5 and LDL < 5.0 or risk factors do not benefit from statin therapy – there’s no reduction in mortality, the NNT is very high for preventing a non-fatal MI or stroke.
6. I have not discussed too much about antihypertensives but we know >160 should have one recommended as higher is unacceptable risk. Follows similar thought train with NNT and NNH for intermediate risk.