Article Text


Definition of acute coronary syndrome
  1. I N Findlay1,
  2. A D Cunningham2
  1. 1Royal Alexandra Hospital, Paisley, UK
  2. 2National Clinical Audit Support Programme, NHS Information Authority, Tavistock House, London, UK
  1. Correspondence to:
    Dr Iain N Findlay
    Royal Alexandra Hospital, Corsebar Road, Paisley, PA2 9PN, UK;

Statistics from

The British Cardiac Society (BCS) Working Group on the definition of acute myocardial infarction1 presents a pragmatic and practical nomenclature for acute coronary syndromes (ACS) and sets definite cut off points for troponin T (TnT) and troponin I (TnI).

We reviewed retrospectively the relation between TnT and creatine kinase (CK) in 978 admissions to our cardiac care unit (over a three year period) who survived to discharge. Criteria for inclusion were simple—that the final discharge diagnosis be recorded, and that patients survive to discharge and had TnT and a peak CK recorded. No attempt was made to correct for sampling time. Our aim was to establish a cut off value for TnT to be equivalent to twice our upper limit of normal for CK (2 × 190  =  380 IU). The mean TnT was 1.09 ng/ml and the mean CK was 433 IU. The correlation between the two, while significant, was not good enough to be clinically useful (fig 1). CK could range from 200–1000 IU for a TnT of 1 ng/ml. A TnT of 0.75 ng/ml was the best analogue of a CK of 380 IU.

Figure 1

 Creatine kinase (CK) versus troponin T (TnT) plotted on a log scale.

Sensitivity, specificity, positive and negative predictive value, and accuracy are given across a range of CK and TnT values in table 1. Receiver operator curves were derived for the sensitivity, specificity, and accuracy for CK and TnT for the clinician discharge diagnosis of acute myocardial infarction (fig 2).

Table 1

 Sensitivity, specificity, positive and negative predictive value, and accuracy across a range of CK and TnT values

Figure 2

 Receiver operator curves (ROC) for CK and TnT for the identification of acute myocardial infarction as define by the discharging clinician.

The data were re-analysed and restricted to survivors across the spectrum of chest pain, excluding conditions such as atrial fibrillation, etc. As expected the specificity of both CK and TnT fell but the best value of TnT remained 0.75 ng/ml as an analogue of a CK value of 380 IU. The cut off value for TnT for a local diagnosis of acute myocardial infarction will probably vary between units. As it is important that we have consistency we are happy to accept a TnT value of 1 ng/ml as the diagnostic cut off for acute myocardial infarction.

As a result of this analysis it became apparent that the diagnostic nomenclature that we were using was unclear to our general medical and general practitioner (GP) colleagues. Clinical coders also commented on the inconsistent terminology and struggled to code ACS with a positive troponin.

General practitioners and rehabilitation staff reported particular difficulties and were often unclear of the patient’s diagnosis. GPs were unable to settle on a common Read code for ACS with positive troponin. With this in mind we have proposed the following nomenclature that attempts to link ICD-10 and Read codes, pending an official reclassification. The proposed changes take into account the mapping carried out by NHS Information Authority Clinical Terminology Browser (v1.04).

We propose an additional term to the BCS classification—that of “ACS unspecified”. This would be in keeping with other coding systems that recognise that coding can be imprecise. We propose that the established ICD-10 code for unstable angina (I200), as well as being used for unstable angina, can be used as a code for ACS which is not otherwise specified at discharge.

We propose that the ICD-10 code I200 be expanded with the extension “TN” to code for the BCS term “ACS with unstable angina” where troponin is negative (I200TN), and with the extension “TP” for the BCS term “ACS with myocyte necrosis” where troponin is positive (I200TP). This latter code has the attraction of mapping readily to the Read term “microinfarction of the heart” (Read code G31y1) occurring in the circumstances of other acute and subacute ischaemic heart disease. The extension “AB” could be added for those acute infarcts aborted by intervention or thrombolytic therapy (I200AB) (table 2).

Table 2

 A proposed terminology and mapping for acute coronary syndromes

These codes map readily to Read codes and will allow primary care to code in a uniform manner. National returns to MINAP will have a more detailed coding system that can be mapped to any new international codes.

The BCS proposed term “ACS with clinical MI” could be regarded as a parent term equivalent and mapped to the parent ICD-10 code for acute myocardial infarction I21 and parent Read code G30 for acute myocardial infarction (table 2). More detail as to site and nature of infarct is obtained by drilling down the coding hierarchy. Note that more clinically descriptive options are provided by the Read code system. It is likely that these more accurately reflect what is described in current hospital discharge letters because of the limitation of ICD-10 codes to describe acute myocardial infarction. If a local unit wishes to code in a more detailed fashion for audit purposes it should consider using the Read terms. This does not preclude hospital coders using ICD-10 codes for central returns.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Miscellanea
    BMJ Publishing Group Ltd and British Cardiovascular Society
  • Miscellanea
    BMJ Publishing Group Ltd and British Cardiovascular Society