Assessing the implications of implementing the NICE guideline 95 for evaluation of stable chest pain of recent onset: a single centre experience

Introduction

The accurate evaluation of patients with recent onset stable chest pain of cardiac origin (i.e. angina) is important and represents a substantial clinical workload in cardiology, acute medicine, emergency medicine and primary care.

When the diagnosis of angina is not clear on clinical grounds alone or when objective clinical evidence to quantify the severity of symptoms is required, it has been the common practice to use supplementary investigations in conjunction with the clinical history.

The practice in Scottish hospitals for the above purpose has been, and is still is, largely guided by the recommendations set by SIGN guideline 96 (http://www.sign.ac.uk/pdf/sign96.pdf), drawn in line with the Recommendations of the Task Force of the European Society of Cardiology ² and the NICE guideline 73 (http://www.nice.org.uk/nicemedia/pdf/TA073guidance.pdf). This guideline states that following careful clinical evaluation, a baseline electrocardiogram (ECG) and an additional non-invasive screening test such as an Exercise Tolerance Test (ETT) be offered. According to this guideline, patients unable to undergo exercise tolerance testing or who have pre-existing electrocardiographic abnormalities should be considered for myocardial perfusion scintigraphy. It also recognises that high-risk patients should be evaluated with coronary angiography but only after the non-invasive tests, or when diagnostic doubt exists. Alternative tests such as stress echocardiography, magnetic resonance perfusion imaging or multi-slice computerised tomography (CT) may be used where suitably trained staff performs these tests, recognising that these are still not part of routine practice of NHS Scotland at the time of issue of the guideline in 2007.

The NICE guideline 95 (issued in March 2010, applicable to England and Wales), represents a change in this traditional approach to investigation of such patients. The main change is that the choice of investigations can be decided by considering the presence or absence of certain high-risk clinical features demonstrated to indicate the presence of significant coronary artery disease (CAD), obviating the mandatory need for screening tests such as exercise tolerance tests in majority of patients. It has been demonstrated that the presence of significant coronary artery disease may be predicted according to the typicality of presenting symptoms (Diamond and Forrester criteria ³ ), age, sex, risk factors and ECG findings (Please see Table 1 – excerpt from NICE Guideline 95: from http://www.nice.org.uk/nicemedia/live/12947/47938/47938.pdf).

The high-risk group (i.e. calculated likelihood of CAD>61–90%), if clinically appropriate, should then be referred for invasive angiography without any further non-invasive tests as this appears to be most diagnostic and cost-effective first test in this group. In those with intermediate risk (i.e. calculated likelihood of CAD = 30–60%), non-invasive functional testing with either myocardial perfusion scintigraphy with SPECT, stress echocardiography, first-pass contrast enhanced magnetic resonance (MR) perfusion or MR imaging for stress-induced wall motion abnormalities should be considered. If the likelihood of CAD is low (calculated likelihood of CAD = 10–29%) computerised tomography (CT) calcium scoring is recommended (with 64 slices or above). The guidance suggests that patients with a low likelihood of CAD will not need further testing (<10% likelihood).

However, if the proposed strategy by NICE to investigate such patients in England and Wales was rolled into NHS Scotland, the provision of cardiology services may need reorganisation, including the need for provision of services on a scale not readily available in Scotland. Therefore, we wanted to assess the contemporary practice of chest pain assessment, using the data of a cohort of patients attending a Rapid Access Chest Pain Clinic (RACPC) in a typical Scottish Hospital, at a time when the NICE guideline 95 was issued, to evaluate the likely scale of change in practices and services that may be required especially in RACPCs, if the SIGN guidelines were to change in future, keeping in line with the NICE Guideline 95, to provide an equity of service across United Kingdom.

Methods

Retrospective analysis of case notes was performed of all patients (n = 96) evaluated with chest pain of recent onset in the RACPC at Ninewells Hospital for two consecutive calendar months (January, February 2010).

Probability of presence of significant CAD was calculated according to presenting symptoms, age, sex, risk factors and ECG findings as per the new NICE guidelines (Pre-test likelihood (PTL) calculator (see Table1)). Investigations actually performed (with their outcomes) and investigations that would have been considered if the new NICE guidance was followed were considered in the analysis. The numbers of patients offered tests (6 types of tests/outcomes as shown in Figure 3) under the current practice and according to NICE proposals in our population were then tabulated in a 2 × 6 contingency table and assessed by a chi-square test of independence.

Results

A total of 96 individuals (of whom 84 patients were never known to have previous clinical CAD), were studied during this period. The demographic features revealed an equal distribution between the sexes (48 males, 48 females), and a normal distribution of age characteristics. The modal distribution was in the 61- to 70-year age group.

Pre-test likelihood (PTL) of presence of significant CAD in all comers was established (Figure 1) as per Table 1 from NICE guideline 95. Modal distribution of likelihood of CAD in patients was in the 30–60% group for females and in the 61–90% group for males. Those who belonged to low-risk categories (PTL <30%) were small (% of total: 3.6% for males and 13.1%f or females) amongst those not previously known to have CAD.

The first investigation offered currently in the RACPC was then assessed (see Figure 2). For most patients, this was an exercise tolerance test (87.6% of all patients, including those deferred to have an ETT at a later date). This was also applicable across all classes of pre-test probability of those not known to have previous CAD (data not shown). Functional imaging (5.2%) or coronary angiography (0%) was not often used as first line diagnostic tests.

Considering the outcomes of the tests, in particular that of the ETTs (as shown in Figure 2), the exercise tolerance tests were positive only in a small minority (4.2%) whereas 24% patients had ETTs that were inconclusive. It is important to review these figures in light of the fact that this cohort of patients had 54.2% of its subjects with a high pre-test likelihood (PTL>60%) of significant CAD (see Figure 1). Even when those who were known to have previous clinical CAD were studied in isolation, only 16.7% had a positive ETT and 41.7% had a negative ETT leading to clinical ‘reassurance’ (data not shown).

Applying the criteria set out in the new NICE guidelines, all patients we studied would have required the following initial tests following clinical assessment: Coronary angiography (42.7%), Functional imaging (29.2%), ETTs (13.5%), CT calcium scoring (8.3%) or no tests (6.3%) (see Figure 3).

This would have meant substantial changes of the use of initial test to be offered to our cohort of patients: there would have been a 74.1% reduction of the offer of ETTs and a 42.7% increases in the offer of coronary angiography, a 24.0% increase in the offer of functional imaging and a 8.3% increase of the offer of CT calcium scoring tests in our study population (see Figure 3). The numbers of patients offered tests under the current practice against that according to NICE proposals in our population was statistically highly significantly different when assessed by chi-square test of independence (χ ²  = 124.0, df = 5, p < 0.0001).

Discussion

The assessment of chest pain of recent onset, with a view for supporting or refuting presence of significant CAD as an aetiology continues to largely involve an ETT as an initial test in our study population. This is likely to reflect the contemporary practice set out to meet standards of the National service framework for CAD, in many hospitals with RACPCs in United Kingdom, prior to the NICE guideline 95 and the Scottish hospitals adhering to the existing SIGN guidelines.

Our study suggests that if the NICE guideline 95 were to be implemented across United Kingdom, the proportion of patients required to be offered specialist tests after initial clinical assessment (coronary angiography (42.7%), functional imaging (29.2%) and CT calcium scoring (8.3%) following initial clinical assessment) would be substantially different from current practice. This would have major implications on the need of expertise necessary at first contact and therefore need to re-organise the process of appropriate and prompt referrals to these tests offered by specialists. We do acknowledge the limitations of our study due to small sample size, limits of time and retrospective case analysis, but it still appears that the application of the new NICE guidelines for the study population would mean that significant changes in the application of first line investigations is needed and this result is very highly unlikely to have arisen by chance (p < 0.0001).

We also have demonstrated that when the current SIGN guideline driven strategy is used, the proportions of those with a clear positive or negative outcome on an exercise tolerance test appear to be small, even in the population categorised to be in the highest pre-test likelihood according to the new NICE guidelines, highlighting the difficulty of using ETTs as a diagnostic tool in patients who are not known to have previous coronary disease.

In their contemporary review, Fox and Mclean ⁴ suggest that application of the NICE guideline would lead to a 20% increase in invasive angiography and a 42% increase in non-invasive imaging from the rapid-access chest pain clinics. We note, albeit of a small sample size, that our study population needed a 42.7% increase in offer of invasive angiography and 24.0% increase in offer of non-invasive imaging as the first line tests. It is possible that regional and seasonal variation of study populations with different incidence and prevalence of CAD can affect the actual percentage increase of the services required at specific RACPCs. Therefore we believe data from much larger studies are required to extrapolate the data nationally to calculate the number of tests and specialist services that may be required in first contact clinics such as RACPCs in Scotland.

However, it should be noted that the exact degree of increased workload for specialist cardiology services as a result of the increased demand for these tests as a first line investigation cannot be directly inferred from our data, as one may argue that these tests may have to be accessed by the same cohort of patients at a later stage of their disease process in any case (albeit experiencing morbidity and mortality whilst waiting to access these specialist services).

We also observe that the NICE guideline 95 has examined the sensitivity and specificity of tests against historic studies with a ‘gold standard’ of angiographically demonstrated CAD ⁵ but this can pose a number of problems: It is known that there is a high prevalence of non-obstructive and functionally unimportant CAD and that the assessment of severity of CAD by conventional angiography even by highly skilled cardiologists can vary. ⁶ It is also known that this can lead to difference of clinical decisions and outcomes. ⁷ Thus it remains to be seen if the dataset from such studies were derived or calibrated with objective physiological measures of coronary blood flow in future (e.g. with use of techniques such as fractional flow reserve (FFR) estimation with pressure wire during coronary angiography) or anatomic visualisation with intra-vascular ultrasound (IVUS), whether we would need to re-calculate the probabilities of presence of significant CAD and also revisit the issue of the investigations likely to detect them in the most cost effective way in a RACPC setting.

Conclusions

Substantial changes to the process of referral, assessment and investigation of patients with stable chest pain of recent onset may be required if the recommendations made in the NICE guideline 95 is universally adapted by hospitals in Scotland. This may require the need to re-organise the services offered by the current rapid access chest pain clinics, to offer more definitive and specialist investigations rather than exercise tolerance tests as first line tests, following initial clinical assessment. Further studies are needed to assess regional variations and evaluate the absolute increase in workload that may be required in the specialist services within cardiology to comply with these recommendations.