detected, then early coronary angiography and interven- tion may improve outcome in these high risk patients. We are indebted to Dr W. Hubbard, Dr R. Thomas ...
European Heart Journal (1998) 19, 240–249
Long-term prognosis in unstable angina The importance of early risk stratification using continuous ST segment monitoring D. J. Patel, C. J. Knight, D. R. Holdright, D. Mulcahy, D. Clarke, C. Wright, H. Purcell and K. M. Fox Department of Cardiology, Royal Brompton Hospital, London, U.K.
Aims To assess the ability of clinical characteristics, admission ECG and continuous ST segment monitoring in determining long-term prognosis in unstable angina. Methods Two hundred and twelve patients with unstable angina (mean age 59 years), presenting within 24 h of an acute episode of angina were recruited at three hospitals and treated with standardized medical therapy. All patients kept chest pain charts and underwent ST segment monitoring for 48 h. The occurrence of death, myocardial infarction, and need for revascularization was assessed over a median follow-up of 2·6 years. Results The risk of death of myocardial infarction was greatest in the first 6–8 weeks after admission. Admission ECG ST depression and the presence of transient ischaemia predicted increased risk of subsequent death or myocardial
Introduction The varied pathophysiology and outcome of patients admitted with a clinical diagnosis of unstable angina[1–3] imposes a need for risk stratification of patients to allow targeting of angiography and revascularization. In the early phase following admission, this may best be achieved on the basis of the clinical and non-invasive screening data available to the clinician. The prognostic importance of refractory angina with significant ECG changes is well recognized[1]; however, it is less clear whether recurrent chest pain carries the same prognostic importance as the presence of transient ischaemia detected by continuous ST segment monitoring. It is also not known whether the determinants of short-term prognosis in unstable angina extend to prognostic value over the long term. Revision submitted 20 May 1997, and accepted 9 June 1997. Correspondence: Dr Deven J. Patel, Department of Cardiology, Harefield Hospital, Hill End Road, Harefield, Middlesex UB9 6JH, U.K. 0195-668X/98/020240+10 $18.00 hj970586
infarction, whereas a normal ECG predicted a good prognosis. In 14 patients, ST segment monitoring provided the only evidence of recurrent ischaemia, and 72% of this group suffered an adverse event. Transient ischaemia and a history of hypertension were the most powerful independent predictors of death or myocardial infarction. Conclusions Adverse events in unstable angina occur early after admission and can be predicted by clinical and ECG characteristics, and by the presence of transient ischaemia during ST segment monitoring. Risk stratification by these simple assessments can identify patients with unstable angina at high risk. (Eur Heart J 1998; 19: 240–249) Key Words: Unstable angina, prognosis, transient ischaemia, ST segment monitoring.
The purpose of this investigation was to examine the prognostic significance of simple clinical and electrocardiographic variables, including the results of continuous ST segment monitoring, in unselected patients admitted with a clinical diagnosis of unstable angina. We have previously reported[4] the in-hospital outcome of unselected patients with unstable angina according to these variables, and we now extend these observations over long-term (median 2·6 years) follow-up.
Methods The detailed methods have been described previously[4,5]. In brief, subjects were recruited from three district general hospitals between 1990–1992 and all results were analysed centrally at the Royal Brompton Hospital. The study received approval from the Ethics committees of all participating centres and all subjects were required to give informed written consent prior to enrolment. ? 1998 The European Society of Cardiology
Long-term prognosis in unstable angina
Patients Subjects aged 30–75 years, presenting within 24 h of an acute episode of anginal pain, requiring emergency admission to the coronary care unit, were recruited. A diagnosis of unstable angina was made on the history and description of chest pain. Patients were included with new onset angina (2 fold rise above baseline), or development of new Q waves on serial ECGs. These subjects have been excluded from this analysis; the remaining 212 subjects with unstable angina form the basis of this report. Patients were followed-up at 6 weeks, 3 and 6 months after hospital discharge, and final follow-up was performed by telephone contact and by review of case notes.
Chest pain Detailed chest pain charts were kept by all patients during the period of ST segment monitoring. For each episode of chest pain, the time and duration were recorded as well as the severity, on a scale of 1 (mild) to 10 (severe). Recurrent pain was defined as the development of pain 24 h after optimal medical therapy; severe pain as any pain of greater than 4 on the scale within the 48 h; and prolonged chest pain as a total duration of painful episodes of >6 min over the 48 h period.
241
blinded to the clinical outcome and the findings of ST segment monitoring. The presence of ST segment depression of §0·02 mV was noted, as were inverted/ biphasic T waves, and the presence of any conduction abnormalities. Differences between observers were mediated by consensus or by a third observer if consensus could not be reached.
ST segment monitoring After enrolment and treatment optimization, patients underwent 48 h of continuous ST segment monitoring using frequency modulated Oxford Medilog MR35 dual channel recorders with a frequency response 0·05–40 Hz. Analysis of the tapes was performed at the Royal Brompton Hospital using an Oxford MA20 analyser blinded to clinical outcome. All potential episodes of ST segment depression were identified by visual review at 60–120 times normal speed, which were then printed and measured. The automatic ST trend was also examined for any further possible episodes and these were confirmed by measurement of the printed ECG trace. Significant ST segment depression was defined as §0·1 mV planar or downsloping ST segment depression from baseline at 80 ms after the J point lasting for more than 60 s. ST elevation was defined as the development of §0·2 mV from the J point. Each episode had to be separated by a return of the ST segment to baseline for at least 1 min to be counted as a discrete episode. Changes in the T wave vector in the absence of ST segment changes were not recorded or included in the analysis. The frequency and duration of episodes, and their association with chest pain was also noted. The recruiting centres were not informed of the results of the ST segment monitoring in order to prevent any referral bias.
Outcome The primary definition of adverse outcome was the occurrence of cardiac death or non-fatal myocardial infarction. Cause of death was determined from the clinical circumstances of the event, the case notes of the hospital where the patient had died, and from postmortem data where available. Myocardial infarction was defined as development of prolonged chest pain with either a greater than two-fold rise in cardiac enzymes above the upper limit of normal, or the development of new Q waves on serial ECGs. A secondary measure of unfavourable outcome was also defined to include cardiac death, non-fatal myocardial infarction as well as the development of severe angina requiring revascularization, either as an emergency or electively.
Resting ECG
Statistical analysis
A 12-lead resting ECG was recorded at admission for all patients. Two observers analysed the ECG and were
All data analysis was performed using the Statistical Package for Social Sciences (SPSS Rel. 6.0) software. Eur Heart J, Vol. 19, February 1998
242
D. J. Patel et al.
Table 1
Clinical characteristics
Clinical characteristics (n=212) Mean age (years) Age >65 years Sex: Male Female Risk factors Smoking history Family history Diabetes Hypertension Cardiac history Previous MI Previous stable angina New onset angina Post MI unstable angina
Table 2 Findings of admission ECG and ST segment monitoring 58·9 61 169 43
(30–77) (28·8%) (80%) (20%)
175 123 14 73
(82·5%) (58·0%) (6·6%) (34·4%)
91 (42·9%) 121 (57·1%) 90 (42·5%) 1
Admission ECG Normal Isolated T wave inversion ST segment depression Old Q waves Other abnormality Chest pain after admission Recurrent pain after 24 h Severe pain (>3/10) in the 48 h Pain duration >60 min/48 h ST segment monitoring Patients with TMI No of TMI episodes Patients with §30 min TMI24 h Patients with §60 min TMI/24 h
61 69 59 23 5
(28·8%) (32·5%) (27·8%) (10·8%) (2·4%)
72 (34%) 90 (42·5%) 19 (9%) 32 (15·1%) 132 (Silent 104) 16 (7·5%) 9 (4·2%)
MI=myocardial infarction. TMI=transient myocardial ischaemia.
Continuous variables are expressed as the mean&1 standard deviation or as median values and range. Two sided P values of significance are quoted throughout with P65 years (61) Males (169) Smoking history (175) Hypertension (73) Diabetes (14) Family history (123) Previous MI (91) New onset angina (90) Recurrent pain (72) Pain >60 min/48 h (19) Severe pain (90)
Any adverse event
No. of events
Hazard ratio
95% CI
P
No. of events
Hazard ratio
95% CI
13 21 21 16 3 12 17 12 11 2 16
2·53 0·90 0·70 3·10 1·83 0·57 2·35 0·90 1·38 0·83 2·08
1·19–5·39 0·36–2·22 0·28–1·73 1·44–6·68 0·55–6·06 0·27–1·21 1·07–5·12 0·42–1·92 0·64–2·97 0·20–3·50 0·97–4·48
0·016
29 87 87 36 9 60 50 40 42 9 43
1·02 1·53 1·21 1·04 1·54 1·04 1·30 1·21 1·68 1·06 1·03
0·66–1·57 0·90–2·61 0·71–2·06 0·70–1·56 0·78–3·06 0·70–1·54 0·89–1·92 0·82–1·82 1·13–2·49 0·54–2·10 0·70–1·53
0·004 0·032
0·073
P
0·010
CI=confidence intervals; MI=myocardial infarction. Eur Heart J, Vol. 19, February 1998
244
D. J. Patel et al.
Table 4
Univariate predictors of outcome from the resting ECG and ST segment monitoring Death or MI
ECG (n) Normal ECG (61) T wave changes (69) ST depression (59) ST segment monitoring (n) Transient ischaemia (32) TIMI §30 min/24 h (16) TIMI §60 min/24 h (9)
Any adverse event
No. of events
Hazard ratio
95% CI
P
No. of events
Hazard ratio
95% CI
P
3 10 12
0·29 1·33 2·31
0·09–0·96 0·63–2·83 1·08–4·94
0·042
0·47 1·38 1·91
0·29–0·77 0·94–2·03 1·28–2·87
0·003
0·03
20 36 36
10 6 5
4·06 2·67 8·55
1·86–8·88 0·98–7·24 3·13–23·3
