Key points are not available for this paper at this time.
Design
Editorial
Answers are grounded in this paper and cite their sources.
This commentary reviews the historical evolution of the definition of myocardial infarction, highlighting the shift from clinical and ECG-based criteria to biomarker-driven universal definitions and the challenges this poses for epidemiology.
James Herrick described non-fatal myocardial infarction a century ago, but attracted little attention until he published again with illustrations of electrocardiograms (ECGs).1 In subsequent decades, confirmation of myocardial infarction in the living necessitated ECG evidence. Newer definitions, based predominantly on recent biomarkers, would make diagnosis difficult without a laboratory. Hence, a World Health Organization (WHO) redefinition, the first in at least two decades, incorporating new biomarkers but re-emphasizing the ECG.2 Classic descriptions were of severe archetypal cases. ‘Clinical judgement’ was used on milder, or atypical cases, or missing data. Myocardial infarction merges into acute coronary syndromes: sudden death, unstable angina, heart failure and chronic coronary heart disease. To compare event rates and case fatalities you need definitions, not descriptions. In 1970, I implemented a myocardial infarction register in the London Borough of Tower Hamlets, part of a WHO European Office collaboration. The pre-existing WHO definition was largely descriptive, with anomalies and ambiguities. Test case histories that I initiated as Rapporteur demonstrated differing interpretations. I tightened the definitions.3 ‘Definites’ were defined, but cases with equivocal or missing data usually became ‘possibles’. Lumped together for publication this increased rates but decreased case fatality.4 Fatal cases involved any post-mortem results complementing or replacing living criteria, but this procedure was often not done. Manipulation of fatal and non-fatal criteria produced differing answers to ‘How often? How fatal?’5 Initiated first to study coronary care, such registers were reborn in the WHO MONICA Project in the 1980s, to monitor trends in coronary heart disease over 10 years in defined populations—possible only in those with the necessary medical resources.6 Again as Rapporteur, I worked with a small group updating the definitions. We incorporated the quantitative Minnesota code for the ECG, which necessitated specialized training, but could not eliminate observer variation.7 Serological cardiac enzyme tests then lacked specificity and sensitivity, and were not standardized as to laboratory methods or how ‘normal’ values were derived. I ran a quality control centre from Dundee and distributed test case histories, followed by final codings, but re-encountered clinical judgement—one centre reported that Dundee was ‘not correct’ as their clinician (criteria not stated), disagreed. MONICA’s first collaborative paper on coronary event rates in 19948 showed correlation for the first time between non-fatal and fatal event rates across populations, but argued that non-fatal events in women in low incidence countries were being missed. It demonstrated inconsistency between local diagnoses, as expressed in ICD codes, and MONICA’s standard categories, a worse problem for fatal than for non-fatal coronary events. Through this publication, MONICA’s diagnostic criteria and definitions appeared in a clinical journal.8 Previously, numerous clinical papers on myocardial infarction had claimed use of ‘WHO criteria’ without quoting a source. Improbable that they used Minnesota coding, most seemed to equate the criteria with ‘two out of three’ of chest pain, elevated cardiac enzymes and ECG progression, but this was over-simplistic. To many cardiologists it was hurtful to find that epidemiologists and the WHO had been defining ‘their’ pre-eminent diagnosis. Quality control and evidence-based medicine were replacing clinical judgement. Explicit criteria were needed not only for clinical trials but also for reimbursement of hospital costs. Emerging biomarkers, the troponins, were sensitive and specific. Commercial and personal interests, and paid consultancies were involved, but myocardial infarction really did need redefining. Unlike MONICA, however, the emphasis was predominantly on patients who arrived alive in a hospital fully equipped for specialist care. Fatal cases were not of interest. The first attempted redefinition, entitled a consensus between the European Society of Cardiology and the American College of Cardiology, was published in 2000. Troponins featured overwhelmingly.9 The derivation had not been transparent. Although invited to, and attending the launch conference and named as a consensus participant in the publication, the resulting ‘redefinition’ took me by surprise as I had seen a vaguely worded conference report, and nothing thereafter. I published my dissent with a critique—among other things the redefinition was descriptive, not a definition, with no room for missing data.10 Subsequently, an American Heart Association Scientific Statement ‘Case definitions for acute coronary heart disease in epidemiology and clinical research studies’,11 in which I did participate, was published in Circulation in 2003. Ongoing objections (I was not alone) from earlier were considered at meetings to which I was not invited. A ‘Universal definition of myocardial infarction’12 with the same lead authors as the 2000 consensus redefinition, but many more sponsors, appeared in cardiology journals in 2007. While the earlier paper9 misquoted WHO criteria but referenced them, this one is more positive about them but fails to list any potential sources3,4,7,8 in its 87 references; also confusing prevalence with incidence.12 The new WHO document (with some shared authors) incorporates much of this ‘universal definition’, but emphasizes, in addition, those unable to measure troponins.2 Paradoxically, therefore, ‘definitions for epidemiology’ appeared in a cardiology journal, but the so-called ‘universal’ definition, when extended to the world outside the troponin-measuring ‘universe’, now appears in this epidemiology journal. The rash of new definitions follows the impact of new technology. When the dust settles, the diagnosis of non-fatal myocardial infarction will have been improved and standardized, but the technology is still advancing. In first-generation assays, any tiny blip in troponin (otherwise usually zero) was controversially taken to diagnose a full myocardial infarct, but new generations of high sensitivity tests give measurable values in a large proportion of the apparently healthy population. With the ECG still the diagnostic tool in many countries now experiencing the cardiovascular disease epidemic, the challenge for reagent manufacturers is to produce a bedside troponin test as cheap and simple as the ECG to administer. A paradoxical outcome could be a reversal of the epidemiology of coronary heart disease of the last century with non-fatal myocardial infarction rates in different populations more reliably assessed than coronary mortality rates, for which troponin has contributed little. I would be reassured to know that newer definitions of myocardial infarction are sufficiently watertight to lead to comprehensive algorithms, as I attempted for MONICA 25 years ago.13 Each and every permutation of variables, including missing values, should lead to one diagnostic category, and one only, with minimal need for clinical judgement and therefore an unambiguous ICD code. Conflict of interest: None declared.
No takes yet. Share an insight, caveat, or question.
Hugh Tunstall‐Pedoe (2010) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: