Clinical (C0205210) Predictors of In-Hospital Prognosis (C0033325~C0220901) in Unstable Angina (C0002965) : ECLA 3

O. Bazzino, MD, FACC, R. Daz, MD, C. Tajer, MD, C. Paviotti, MD, E. Mele, MD, M. Trivi, MD, A. Piombo, MD, A. Hirschson Prado, MD, E. Paolasso, MD, Servicio de Cardiologa, Hospital Italiano, Buenos Aires, Argentina; for the ECLA Collaborative Group.

Am Heart J13721999

1372

Abstract

Objectives (C0018017) : Because of recent changes (C0025320~C0392747~C0443172) in the treatment (C0039798~C0087111) of unstable angina (C0002965) , we wanted to reassess the short-term prognostic value (C0042295~C0401806) of clinical and echocardiographic variables (C0439828) .
Methods (C0025663~C0025664) : This was an observational , prospective study (C0033522) that included 1038 nonselected consecutive patients (C0030705) admitted to coronary care units (C0010066) for unstable angina (C0002965) .
Results (C0332294) : Baseline (C0168634) characteristics were age (C0001779~C0001792~C0001811~C0002555~C0037125~C0521993) 60.18 +-16 years (C0439234~C0439508) , history (C0019664~C0019665~C0262512~C0262926) of prior myocardial infarction (C0027051) in 336 patients (C0030705) ( 32 % ) , and a history (C0019664~C0019665~C0262512~C0262926) of previous angina (C0002962) in 817 patients (C0030705) ( 78.7 % ) .
Angina (C0002962) during the 48 hours (C0439227) before admission (C0457453) was observed in 1004 patients (C0030705) ( 96.7 % ) and ST-segment changes (C0025320~C0392747~C0443172) on admission electrocardiogram (C0013798) occurred in 385 patients (C0030705) ( 37 % ) .
In-hospital treatment (C0039798~C0087111) consisted of nitrates (C0028125) in 81.4 % of patients (C0030705) , aspirin (C0004057) in 88.6 % , b-blockers in 71 % , intravenous heparin (C0354566) in 34.5 % , subcutaneous heparin (C0353681) in 23 % , and angioplasty (C0162577) or coronary artery bypass (C0010055) grafting in 25.1 % .
After admission (C0457453) , angina (C0002962) occurred in 443 patients (C0030705) ( 40.8 % ) , refractory angina (C0002962) in 223 patients (C0030705) ( 21.5 % ) , and death (C0011065~C0220816) or myocardial infarction (C0027051) in 84 patients (C0030705) ( 8.1 % ) .
At admission (C0457453) , the independent (C0332291) predictors of myocardial infarction (C0027051) or death (C0011065~C0220816) identified by multivariate logistic regression analysis (C0034980) were ST-segment depression ( odds ratio (C0028873) [ OR (C0134121) ] 2.13 , 95 % confidence interval (C0009667) [ CI (C0009968~C0056599) ] 1.23 to 3.68 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .006 ) , prior angina (C0002962) ( OR (C0134121) 2.23 , 95 % CI (C0009968~C0056599) 0.98 to 5.05 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .05 ) , number (C0237753~C0449788) of episodes (C0277793~C0332189) of angina (C0002962) within the previous 48 hours (C0439227) ( OR (C0134121) 1.63 , 95 % CI (C0009968~C0056599) 0.98 to 2.70 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .05 ) , and history (C0019664~C0019665~C0262512~C0262926) of smoking (C0037366~C0037369~C0337664) ( OR (C0134121) 0.69 , 95 % CI (C0009968~C0056599) 0.56 to 0.85 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .
004 ) .
Age greater than 65 years (C0439234~C0439508) ( OR (C0134121) 1.49 , 95 % CI (C0009968~C0056599) 1.09 to 2.03 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = 0.03 ) was significantly related to in-hospital death (C0277608) .
The area (C0205146) under the receiver operating characteristic curve (C0205134) for application (C0185125) of this model (C0026336) was 0.59 .
Sensitivity (C0036667~C0312418~C0427965~C0518607) was 80 % with a specificity (C0037791) of only 33 % .
Refractory angina (C0002962) after admission (C0457453) showed a strong (C0442821) relation with an adverse short-term (C0443303) outcome .
Conclusions : With current therapy (C0039798~C0087111) , clinical and electrocardiographic variables (C0439828) provide useful information about the short-term (C0443303) outcome of unstable angina (C0002965) .
However , this model (C0026336) has low specificity (C0037791) to identify high-risk patients (C0030705) .
Future (C0016884) studies about the incremental value (C0042295~C0401806) of the new serum markers (C0162491) such as troponin T (C0077404) and C-reactive protein (C0006560~C0201657) to assist in identification (C0020792) of high-risk patients (C0030705) are necessary .

Introduction

Management of unstable angina (C0002965) has become more aggressive to include full medical therapy (C0418981) with anti-thrombotic and anti-ischemic drugs (C0013227~C0278180) , often followed by coronary angiography (C0085532) and revascularization procedures (C0025664~C0184661) .

Much of the published (C0034037) data about the natural history (C0175860) of unstable angina (C0002965) corresponds to selected patients (C0030705) included in randomized clinical trials (C0206034) .
Therefore these patients (C0030705) may not represent the general population (C0032659) admitted to community hospitals (C0020003) .

Unstable angina (C0002965) is also quite heterogeneous , with a wide (C0332464~C0426421) spectrum of clinical manifestations (C0205319) and prognostic categories .
in addition (C0332287) , most of the data on the evolution (C0015219) and prognosis (C0033325~C0220901) of unstable angina (C0002965) were obtained before the development (C0243107) of new drugs (C0013227~C0278180) and interventional procedures (C0025664~C0184661) for management of this disease (C0012634~C0028432~C0263648~C0274294) .

This multicenter study (C0026739~C0282439) enrolled a large (C0205164~C0426415~C0443228) cohort of patients (C0030705) admitted with unstable angina (C0002965) to identify the *nonlinear*: clinical markers (C0008963) clinical and electrocardiographic markers associated with in-hospital death (C0277608) and myocardial infarction (C0027051) , in the context (C0449255) of a nonselected population (C0032659) admitted consecutively to community hospitals (C0020003) and treated according to current therapeutic (C0039796~C0302350) strategies .

Methods (C0025663~C0025664)

Study Design

This was a multicenter prospective study (C0033522) of consecutive patients (C0030705) admitted to coronary care units (C0010066) with unstable angina (C0002965) .
To select a representative sample (C0441621) and safeguard against selection bias (C0036577) , all consecutive patients (C0030705) admitted to the coronary care unit (C0010066) with a suspicion (C0242114) of an acute ischemic syndrome (C0039082) were included in the database (C0242356) .
If the admission diagnosis (C0011900~C0011901) turned out to be Q-wave (C0429089) or non-Q-wave myocardial infarction (C0027051) , or if a nonischemic cause (C0015127) of chest pain (C0008031) became clearly evident , the process of data collection was interrupted and the patient (C0030705) censored from subsequent analysis (C0002778) .
The diagnosis (C0011900~C0011901) of Q-wave (C0429089) or non-Q-wave acute myocardial infarction (C0155626) present at admission (C0457453) was made on the basis (C0004830~C0178499) of the finding (C0243095~C0263541~C0332285) of an elevated total creatine phosphokinase (C0010287) ( CPK (C0010287) ) plasma level (C0441889~C0456079) ( greater than 2-fold increase (C0205217~C0442805~C0442808) above normal value (C0086715~C0220882) ) at entry to the study or at any time (C0040223~C0183941~C0332311~C0392761~C0449243) up to 12 hours (C0439227) thereafter .
for patients (C0030705) prescribed thrombolytic treatment (C0039798~C0087111) or primary coronary angioplasty (C0190211) at entry , the enzymatic criteria (C0243161) were not required and the diagnosis (C0011900~C0011901) was made on the basis (C0004830~C0178499) of the finding (C0243095~C0263541~C0332285) of chest pain (C0008031) lasting more than 30 minutes (C0205165~C0439232) plus > = 1 mm ST elevation (C0429069) in at least 3 precordial or 2 standard electrocardiogram (C0013798) ( ECG (C0013798) ) leads (C0023175~C0181586~C0373667~C0475211) .

Between March 1991 and May l (C0023175~C0181586~C0373667~C0475211) 992 , 1172 consecutive patients (C0030705) with suspected acute ischemic syndrome (C0039082) were prospectively enrolled .
This group (C0439745~C0441833) represented 23 % of the total population (C0032659) admitted to the coronary care units (C0010066) during that period (C0025344~C0439531) of time (C0040223~C0183941~C0332311~C0392761~C0449243) .
One hundred and thirty-four patients (C0030705) ( 11.4 % ) were excluded : 46 patients (C0030705) because of an evolving myocardial infarction (C0027051) unequivocally present at hospital admission (C0184666) and 88 patients (C0030705) because of noncardiac chest pain (C0008031) .
The remaining 1038 patients (C0030705) with unstable angina (C0002965) formed the study population (C0032659) .

Data collected include demographics (C0011298) and baseline (C0168634) clinical variables (C0439828) .
Electrocardiograms (C0013798) were routinely obtained at admission (C0457453) , during subsequent episodes (C0277793~C0332189) of chest pain (C0008031) , and at hospital discharge (C0012621~C0264273) .
Serum (C0229671) total CPK levels (C0441889~C0456079) were determined at admission (C0457453) , every 12 hours (C0439227) during the first 24 hours (C0439227) , and after a new chest pain (C0008031) episode during hospital (C0019994) stay .
Changes (C0025320~C0392747~C0443172) in the ECG (C0013798) were considered positive if there was > = 1 mm (C0126723~C0439200) of ST-segment elevation (C0520886) or depression (C0011570~C0460137) in at least 2 limb leads (C0441100) or 3 precordial leads (C0023175~C0181586~C0373667~C0475211) or symmetric T-wave inversion (C0520888) in at least 3 limbs (C0015385~C0431777) or precordial leads (C0023175~C0181586~C0373667~C0475211) .

The treatment (C0039798~C0087111) received , any coronary intervention , and all the in-hospital events (C0441471) ( angina (C0002962) , acute myocardial infarction (C0155626) , and death (C0011065~C0220816) ) were also recorded .
The investigator (C0035173) who was responsible for the patient (C0030705) was instructed to carry a daily (C0332173) prospective and systematic register of all new chest pain (C0008031) episodes in addition (C0332287) to all the information related to the associated (C0332281) electrocardiographic and enzymatic alterations .
All the ECG tracings (C0442822) were examined by an ECG core lab reader (C0034754) who was blinded to the patient clinical condition (C0009647~C0348080) to validate ST-segment (C0429029) changes > = 1 mm (C0126723~C0439200) , T-wave changes (C0025320~C0392747~C0443172) , left bundle branch block (C0023211) , and left ventricular hypertrophy (C0149721) .
Serial electrocardiograms (C0013798) were compared to evaluate if the changes (C0025320~C0392747~C0443172) were transient or fixed .
The medical treatment (C0039798~C0087111) was entirely left to the discretion of the treating physician (C0031831) except for a strong recommendation (C0034866) to use aspirin (C0004057) in all patients (C0030705) .

End Points (C0324413) and Definition of Events (C0441471)

New episodes (C0277793~C0332189) of ischemic chest pain (C0008031) after admission (C0457453) were classified as ( 1 ) recurrent angina (C0002962) , defined as angina (C0002962) occurring in the absence (C0332197~C0424530) of full medical therapy (C0418981) or ( 2 ) refractory angina (C0002962) , angina (C0002962) occurring while the patient (C0030705) was receiving full medical therapy (C0418981) .

Full medical therapy (C0418981) was defined as the combination (C0205195~C0453882) of aspirin (C0004057) and / or intravenous heparin (C0354566) , nitrates (C0028125) , and b-blockers or calcium antagonists (C0243076) at clinically effective doses (C0018081~C0057911~C0114856~C0231807) .
The adequacy for full medical therapy (C0418981) was left to the decision of the treating physician (C0031831) with suggested criteria (C0243161) of heart rate (C0018810) < 60 beats / min (C0026175~C0240346~C0439232) and systolic blood pressure (C0428880~C0520844) < 120 mm Hg (C0439475) .

Major in-hospital events (C0441471) were new myocardial infarction (C0027051) and death (C0011065~C0220816) from any cause (C0015127) .
Hospital mortality (C0085556) was defined as death (C0011065~C0220816) occurring in the hospital (C0019994) at any moment from arrival through discharge (C0012621~C0264273) .
The criteria (C0243161) for new in-hospital (C0019994) acute myocardial infarction (C0155626) included all of the following (C0332282~C0332283) : ( 1 ) new prolonged episode (C0277793~C0332189) of chest pain (C0008031) after admission (C0457453) ; ( 2 ) elevation (C0439775) of total CPK level (C0441889~C0456079) greater than twice the limit (C0439801) of normal (C0205307~C0439166) , and ( 3 ) the development (C0243107) of new Q-wave (C0429089) or ST-T changes (C0025320~C0392747~C0443172) on the ECG (C0013798) .

Events (C0441471) that occurred during or within 24 hours (C0439227) after an invasive procedure (C0038895~C0348025) ( coronary angiography (C0085532) , coronary angioplasty (C0190211) , or coronary bypass (C0010055) surgery ) were considered procedure (C0025664~C0184661) related .
The remaining events (C0441471) were prospectively classified as spontaneous .

An ad-hoc committee blinded to the clinical variables (C0439828) and the administered treatment (C0039798~C0087111) examined and validated all major events (C0441471) ( acute myocardial infarction (C0155626) or death (C0011065~C0220816) during hospitalization (C0019993) ) with a systematic review (C0282443) of serial electrocardiograms (C0013798) , enzymatic curves (C0205134) , and clinical charts (C0007963~C0178917) of all patients (C0030705) who had major events (C0441471) reported or adjudicated by investigators (C0035173) .

Statistical Analysis (C0002778)

The initially targeted sample (C0441621) of 1000 patients (C0030705) was based on a projected 24 % rate (C0034693) of events (C0441471) ( 16 % refractory angina (C0002962) , 6 % acute myocardial infarction (C0155626) , and 2 % death (C0011065~C0220816) ) .
An a level (C0441889~C0456079) of significance < .05 was considered for all purposes (C0449256) .
Unadjusted odds ratios (C0028873) and 95 % confidence intervals (C0009667) were obtained for selected clinical (C0205210) factors judged to be associated with the outcome of interest (C0123730~C0205103) .
Univariate and forward stepwise multiple logistic regression (C0206031) models were used to evaluate the independent association (C0004083) of variables (C0439828) present at admission (C0457453) to in-hospital (C0019994) acute myocardial infarction (C0155626) or death (C0011065~C0220816) .
From 28 clinical and electrocardiographic variables (C0439828) , those identified of potential (C0237399) significance from the univariate analysis (C0002778) were chosen for forward (C0439780) stepwise selection (C0036576) into the model (C0026336) .
The variables (C0439828) included in the model (C0026336) were age ( with a cutoff point (C0324413) of 65 years (C0439234~C0439508) ) , prior myocardial infarction (C0027051) , prior angina (C0002962) ( angina (C0002962) preceding the 48-hour (C0439586) interval before admission (C0457453) ) , smoking ( prior or current ) , the number (C0237753~C0449788) of angina episodes (C0277793~C0332189) during the last 48 hours (C0439227) before admission (C0457453) ( with a cutoff rate (C0034693) of 3 hours (C0439227) ) , the presence (C0392148~C0392743) of precipitating factors of angina (C0002962) ( anemia (C0002871) , tachyarrythmias , fever (C0015967) , hypertension (C0020538) ) , and the finding (C0243095~C0263541~C0332285) of ST-segment depression (C0520887) on the admission ECG (C0013798) .
Progressive angina (C0340290) was not included because of its significant association (C0004083) with prior angina (C0002962) ( Pearson 's r (C0035253~C0175724~C0332153~C0332575~C0439117~C0439142) = 0.70 ) .

The model (C0026336) was validated with the split-sample technique (C0025664~C0449851) .
A simple random sample (C0150105) of approximately 70 % of the 1038 cases ( the training set (C0036849~C0324444) , n (C0027552~C0028158~C0332125~C0347979~C0439138~C0439225~C0441922) = 728 ) was used to derive the predictive model (C0026336) ; the remaining 30 % sample (C0441621) ( the test (C0039593~C0039597~C0076174~C0154007~C0183885~C0392366~C0430738~C0496924) set , n (C0027552~C0028158~C0332125~C0347979~C0439138~C0439225~C0441922) = 310 ) was used to test it .
The rule (C0205159~C0522637) obtained from the training set (C0036849~C0324444) was applied to each member of the test (C0039593~C0039597~C0076174~C0154007~C0183885~C0392366~C0430738~C0496924) set .
The efficiency (C0013682) of the model (C0026336) was assessed by comparison of the area (C0205146) under the receiver operating characteristic ( ROC ) curves of the training (C0040607~C0220931~C0348006~C0453962) and test samples (C0441621) .

The model (C0026336) was also tested on all 1038 cases to determine its overall predictive ability (C0085732) ( by use (C0042153~C0439224~C0449865) of the area (C0205146) under the ROC curve (C0035787) ) , sensitivity (C0036667~C0312418~C0427965~C0518607) , and specificity (C0037791) in predicting in-hospital death (C0277608) or myocardial infarction (C0027051) .
The goodness of fit (C0036572~C0441548) of the model (C0026336) was assessed by the Hosmer-Lemeshow test (C0039593~C0039597~C0076174~C0154007~C0183885~C0392366~C0430738~C0496924) .

Kaplan-Meier survival curves (C0205134) were generated to calculate the cumulative probability (C0033204) of having a major event (C0441471) develop during hospitalization (C0019993) in the overall population (C0032659) and in the subgroups with and without refractory angina (C0002962) .
Statistical analysis (C0002778) was performed with the SPSS system (C0449913) 7.5 package (C0030176) ( Statistical Package (C0030176) for the Social Sciences (C0037434) , 1997 ) and the ROC Analyzer Program version (C0333052) 6.0 .

Results (C0332294)

Population Demographics (C0011298) and Characteristics

The patients (C0030705) in this study were mostly elderly ( age (C0001779~C0001792~C0001811~C0002555~C0037125~C0521993) 60.18 +-16 years (C0439234~C0439508) ) , 65 % were male , 78 % reported prior angina (C0002962) ( angina (C0002962) preceding the 48 hours (C0439227) interval before admission (C0457453) ) , and 32 % had experienced a prior myocardial infarction (C0027051) ( Table (C0039224) I ) .
The study population (C0032659) was clearly unstable as manifested by the high frequency (C0205212) of resting chest pain (C0008031) ( 78 % ) , prolonged ( > 30 minutes (C0205165~C0439232) ) chest pain (C0008031) ( 29 % ) and the last episode (C0277793~C0332189) of pain (C0030193~C0240564) within the 48 hours (C0439227) preceding admission (C0457453) ( 96.7 % ) .
Angina (C0002962) within 4 weeks (C0439230~C0439506) and 2 weeks (C0439230~C0439506) after infarction (C0021308) ( postinfarction angina (C0002962) ) was observed in 9.7 % and 3.3 % of cases , respectively .
Precipitating factors of unstable angina (C0002965) were not identified in 76 % of the population (C0032659) .
There were relatively few smokers (C0037366~C0037369~C0337664) ( 21.3 % ) , and the most prevalent risk factors (C0035648) were hypertension ( 60.3 % ) and hyperlipidemia (C0020473) ( 43.3 % ) .
The presence (C0392148~C0392743) of heart failure (C0018801) at admission (C0457453) was infrequently reported ( 1.5 % ) .

Left ventricular hypertrophy (C0149721) or complete left bundle branch block (C0264915) precluded the analysis (C0002778) of ST-T changes (C0025320~C0392747~C0443172) in 12.7 % of patients (C0030705) .
ST segment elevation (C0520886) was frequently observed ( 16 % ) and , in contrast (C0110625) , ST-segment depression (C0520887) was relatively infrequent ( 21 % ) .
Only 16 % of the patients (C0030705) had a normal ECG (C0239237) at entry .

Before hospitalization (C0019993) , treatment (C0039798~C0087111) consisted of aspirin (C0004057) in half of the patients (C0030705) ( 50.6 % ) , b-blockers in 31 % , nitrates (C0028125) in 44 % , and calcium antagonists (C0243076) in 42 % .
Combined treatment (C0039798~C0087111) with aspirin (C0004057) , nitrates (C0028125) , and b-blockers or calcium antagonists (C0243076) was administered to 27 % of the study population (C0032659) .

Table II (C0439069~C0439453) depicted the baseline (C0168634) characteristics of patients (C0030705) with and without major complications (C0009566~C0009567~C0392327) during hospitalization (C0019993) .
Patients (C0030705) with complications (C0009566~C0009567~C0392327) were older and had greater frequency (C0042023~C0376249~C0439603) of prior angina (C0002962) , ST-segment depression (C0520887) on admission electrocardiogram (C0013798) , and repetitive angina (C0002962) ( > 3 episodes (C0277793~C0332189) ) during the last 48 hours (C0439227) preceding admission (C0457453) .

Conversely , the history (C0019664~C0019665~C0262512~C0262926) of previous infarction (C0021308) was less frequent among patients (C0030705) without complications (C0009566~C0009567~C0392327) during hospitalization (C0019993) .

In - Hospital (C0019994) Management

Antithrombotic treatment (C0039798~C0087111) was administered in the hospital (C0019994) to 94 % of patients (C0030705) , with heparin (C0019134) in 57 % ( intravenous in 34 % , subcutaneous heparin (C0353681) in 23 % ) , aspirin (C0004057) in 88 % , and ticlopidine (C0040207) in 5 % of the patients (C0030705) .
Heparin (C0019134) was used in 34.7 % and 63.9 % of patients (C0030705) with or without precipitating factors , respectively .
Anti-ischemic therapy (C0039798~C0087111) consisted of the administration (C0001554) of nitrates (C0028125) in 81 % , b-blockers in 71 % , and calcium antagonists (C0243076) in 56 % .

A coronary angiography (C0085532) was performed in 50.2 % of the population (C0032659) .
The most relevant angiographic findings (C0243095~C0263541~C0332285) were normal or nonsignificant stenosis (C0038255) in 5.4 % , 1-vessel disease (C0012634~C0028432~C0263648~C0274294) in 32 % , multivessel disease (C0012634~C0028432~C0263648~C0274294) in 61 % , and left main involvement (C0205428) in 10 % .

A quarter of the population (C0032659) was submitted to revascularization procedures (C0025664~C0184661) ( coronary artery bypass grafting (C0010055) was performed in 14 % and coronary angioplasty (C0190211) in 10.8 % ) .
There was a positive (C0205159) relation between the presence (C0392148~C0392743) and severity (C0237870~C0439793~C0449294) of angina (C0002962) after admission (C0457453) and the need (C0027552~C0028158~C0332125~C0347979~C0439138~C0439225~C0441922) for invasive procedures (C0038895~C0348025) , either diagnostic or therapeutic : 29.5 % and 44.2 % in patients (C0030705) with recurrent and refractory angina (C0002962) , respectively , and 11.1 % in the subgroup without angina (C0002962) after admission (C0457453) ( P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) < .
001 ) .
The indications (C0237000~C0521089) for revascularization procedures (C0025664~C0184661) were angina after admission (C0457453) in 71 % , inducible ischemia (C0022116) in 6 % , angiographic findings (C0243095~C0263541~C0332285) in 21 % , and preference of the patient or his / her physician (C0031831) in the remainder .

Clinical Events (C0441471)

During hospital (C0019994) stay , 40.8 % of the patients (C0030705) had at least 1 new episode (C0277793~C0332189) of cardiac pain (C0030193~C0240564) ( Table III (C0439070) ) .
Recurrent angina (C0002962) was diagnosed in 200 patients (C0030705) ( 19.3 % ) and refractory angina (C0002962) in 223 ( 21.5 % ) patients (C0030705) .

New acute myocardial infarction (C0155626) was diagnosed in 54 patients (C0030705) ( 5.2 % ) , and there were 42 in-hospital deaths (C0277608) ( 4.1 % ) .
The combined end point (C0324413) of death (C0011065~C0220816) and myocardial infarction (C0027051) occurred in 83 patients (C0030705) ( 8.1 % ) .

The incidence (C0021149~C0220856) of new acute myocardial infarction (C0155626) was higher during the first day (C0439228~C0439229~C0439505) and declined thereafter .
On the other hand (C0018563) , cases of death (C0011065~C0220816) were evenly distributed during hospitalization (C0019993) .
The Kaplan-Meier event-free survival (C0242793) curve for in-hospital infarction (C0021308) or death (C0011065~C0220816) is shown in Fig (C0349966) 1 .
Fig (C0349966) 2 shows the in-hospital survival curve (C0205134) of patients (C0030705) free of infarction (C0021308) , death (C0011065~C0220816) , or refractory angina (C0002962) .

The rate (C0034693) of " spontaneous " ( non-procedure-related ) acute myocardial infarction (C0155626) and death (C0011065~C0220816) was 3.1 % and 1.4 % , respectively .
Acute myocardial infarction (C0155626) and death (C0011065~C0220816) related to an interventional procedure rates (C0034693) were 1.0 % and 3.8 % .

Relation Between Clinical (C0205210) Characteristics and Outcome

Univariate analysis (C0002778)

The clinical variables (C0439828) present at admission (C0457453) and associated with an adverse outcome through the hospitalization (C0019993) period are summarized in Table IV (C0348016) .
New acute myocardial infarction (C0155626) or death (C0011065~C0220816) was associated with ST-segment depression (C0520887) ( odds ratio (C0028873) [ OR (C0134121) ] 2.00 , 95 % confidence interval (C0009667) [ CI (C0009968~C0056599) ] 1.20 to 3.40 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .008 ) , the number (C0237753~C0449788) of chest pain (C0008031) episodes within the last 48 hours (C0439227) ( OR (C0134121) 1.84 , 95 % CI (C0009968~C0056599) 1.12 to 3.07 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .01 ) , prior angina (C0002962) ( OR (C0134121) 2.70 , 95 % CI (C0009968~C0056599) 1.34 to 5.57 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .001 ) , and age (C0001779~C0001792~C0001811~C0002555~C0037125~C0521993) > 65 years (C0439234~C0439508) ( OR (C0134121) 1.64 , 95 % CI (C0009968~C0056599) 1.00 to 2.70 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .
01 ) .
A negative trend (C0040833) was noted for smoking history (C0019664~C0019665~C0262512~C0262926) ( OR (C0134121) 0.51 , 95 % CI (C0009968~C0056599) 0.31 to 0.83 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .005 ) , and prior acute myocardial infarction (C0155626) ( OR (C0134121) 0.58 , 95 % CI (C0009968~C0056599) 0.32 to 1.03 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .
039 ) .
Age (C0001779~C0001792~C0001811~C0002555~C0037125~C0521993) ( OR (C0134121) 2.26 , 95 % CI (C0009968~C0056599) 1.06 to 4.69 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .01 ) , and prior angina (C0002962) ( OR (C0134121) 4.9 , 95 % CI (C0009968~C0056599) 1.16 to 12.5 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .009 ) , were the only univariate predictors of in-hospital mortality (C0085556) .

Multivariate analysis (C0026777)

The model (C0026336) was constructed with all the variables (C0439828) present at admission (C0457453) and found significant at the univariate analyses (C0002778) at a level (C0441889~C0456079) of P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) < .
10 .
As shown in Table V (C0015751~C0016129~C0022084~C0042306~C0042584~C0150233~C0182934~C0220828~C0439109~C0439120~C0439132~C0439133~C0439145~C0453884) , the multivariate analysis (C0026777) retained the following independent admission (C0457453) predictors of myocardial infarction (C0027051) or death (C0011065~C0220816) : ischemic ST-segment depression (C0520887) in the ECG (C0013798) upon admission (C0457453) ( OR (C0134121) 2.13 , 95 % CI (C0009968~C0056599) 1.23 to 3.68 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .006 ) , prior angina (C0002962) ( OR (C0134121) 2.23 , 95 % CI (C0009968~C0056599) 0.98 to 5.05 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .05 ) , the number (C0237753~C0449788) of angina episodes (C0277793~C0332189) within the last 48 hours (C0439227) before admission (C0457453) ( OR (C0134121) 1.63 , 95 % CI (C0009968~C0056599) 0.98 to 2.70 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .05 ) , and smoking history (C0019664~C0019665~C0262512~C0262926) ( OR (C0134121) 0.69 , 95 % CI (C0009968~C0056599) 0.56 to 0.85 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .
004 ) .
Age (C0001779~C0001792~C0001811~C0002555~C0037125~C0521993) > 65 years (C0439234~C0439508) ( OR (C0134121) 1.49 , 95 % CI (C0009968~C0056599) 1.09 to 2.03 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .03 ) was significantly related to in-hospital death (C0277608) .

The overall predictive efficiency (C0013682) was similar in the training (C0040607~C0220931~C0348006~C0453962) and test samples (C0441621) ( ROC curves areas (C0205146) 0.57 +-0.05 and 0.53 +-0.04 respectively , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .
29 ) .

On all 1038 patients (C0030705) , the area (C0205146) under the ROC curve (C0035787) for the model (C0026336) was 0.59 +-0.03 .
The sensitivity (C0036667~C0312418~C0427965~C0518607) and specificity (C0037791) of the model (C0026336) were computed .
To ensure at least 80 % sensitivity (C0036667~C0312418~C0427965~C0518607) , a predicted probability (C0033204) cutpoint of 0.06 was required ( ie , if the predicted probability (C0033204) was > = 0.06 , the patient (C0030705) was classified as positive ) .
This cutpoint resulted in a sensitivity (C0036667~C0312418~C0427965~C0518607) of 80 % and a specificity (C0037791) of 33 % .
The Hosmer-Lemeshow goodness-of-fit statistic (C0038215~C0220917) was 3.51 ( degrees (C0439489~C0449286) of frequency (C0042023~C0376249~C0439603) = 7 , P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) = .
83 ) .

Postadmission Angina (C0002962) and Prognosis (C0033325~C0220901)

The occurrence (C0243132) of angina (C0002962) after admission (C0457453) ( recurrent or refractory ) showed a strong (C0442821) univariate relation with the incidence (C0021149~C0220856) of in-hospital (C0019994) acute myocardial infarction (C0155626) or death (C0011065~C0220816) ( 14.4 % vs 3.2 % , or 5.04 , 95 % CI (C0009968~C0056599) 2.86 to 8.87 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) < .0001 ) and death (C0011065~C0220816) alone ( 6.9 % vs 1.1 % , or 6.45 , 95 % CI (C0009968~C0056599) 2.61 to 5.9 ; P (C0016504~C0033452~C0036421~C0078414~C0439115~C0439140~C0449201~C0487890) < .0001 ) ( Table III (C0439070) ) .
Fig (C0349966) 3 shows the Kaplan-Meier infarction-free survival curves (C0205134) in patients (C0030705) with or without refractory angina (C0002962) .
Absence (C0332197~C0424530) of refractory angina (C0002962) from admission (C0457453) to discharge was associated to a better hospital (C0336603) outcome .
This association (C0004083) persisted even when only non-procedure-related ( " spontaneous " ) events (C0441471) were taken into consideration (C0518609) 10.3 % vs 2.3 % of acute myocardial infarction (C0155626) or death (C0011065~C0220816) in patients (C0030705) with or without refractory angina (C0002962) , respectively ( OR (C0134121) 4.40 , 95 % CI (C0009968~C0056599) 2.35 to 8.22 ; Table IV (C0348016) ) .

Discussion

Despite improved treatment (C0039798~C0087111) of unstable angina (C0002965) during the last years (C0439234~C0439508) , there remains a considerable risk (C0035647) of events (C0441471) upon admission (C0457453) to the hospital (C0019994) .
Unstable angina (C0002965) includes a wide (C0332464~C0426421) variety of patients (C0030705) with different prognoses (C0033325~C0220901) ; accordingly , a reliable risk (C0035647) stratification for everyday practice is necessary .
This is a major (C0205164) challenge because most of the short-term events (C0441471) are concentrated in an small number (C0237753~C0449788) of patients (C0030705) .
In addition (C0332287) , to select appropriate treatment (C0039798~C0087111) the risk assessment (C0086930) should be performed as early as possible , ideally during the first hours (C0439227) after admission (C0457453) .

Prognostic Assessment at Hospital Presentation (C0449450)

This study indicates that with current treatment (C0039798~C0087111) the clinical variables (C0439828) continue to be useful short-term (C0443303) predictors at the time (C0040223~C0183941~C0332311~C0392761~C0449243) of hospital admission (C0184666) .

After performing multivariate logistic regression analysis (C0034980) , prior angina (C0002962) , ST-segment depression (C0520887) at the admission ECG (C0013798) , the number (C0237753~C0449788) of pain episodes (C0277793~C0332189) in the last 48 hours (C0439227) preceding admission (C0457453) , smoking history (C0019664~C0019665~C0262512~C0262926) , and increased age (C0001779~C0001792~C0001811~C0002555~C0037125~C0521993) were all independently associated to adverse in-hospital (C0019994) outcome .
However , this model (C0026336) showed low specificity (C0037791) ( 33 % ) .

In this study it was also confirmed that at the time (C0040223~C0183941~C0332311~C0392761~C0449243) of hospital presentation (C0449450) , the severity (C0237870~C0439793~C0449294) of preadmission symptoms (C0178311~C0426409) is an important determinant of in-hospital prognosis (C0033325~C0220901) .
Prior angina (C0002962) and the number (C0237753~C0449788) of episodes (C0277793~C0332189) of angina (C0002962) during the 48 hours (C0439227) preceding admission (C0457453) were independent predictors of an adverse outcome by multivariate logistic regression analysis (C0034980) .
This finding (C0243095~C0263541~C0332285) is a contribution to the elucidation of the controversy regarding the importance of these clinical (C0205210) features as indicators (C0021212) of high risk (C0332167) .
In the past , both a long history (C0019664~C0019665~C0262512~C0262926) of ischemic disease (C0012634~C0028432~C0263648~C0274294) and the lack (C0332268) of preceding symptoms (C0178311~C0426409) have been suggested to identify high-risk (C0332167) subgroups .

Our study confirms that prognosis (C0033325~C0220901) is worse in patients (C0030705) with ST-segment depression (C0520887) on admission ECG (C0013798) .
As shown in previous (C0205156) studies , T-wave changes (C0025320~C0392747~C0443172) , although common , have no such prognostic value (C0042295~C0401806) .
This analysis (C0002778) did not take into account (C0000938) the extension (C0231448~C0332258) and severity (C0237870~C0439793~C0449294) of ST-segment (C0429029) or T-wave changes (C0025320~C0392747~C0443172) and therefore we cannot rule out the possibility of a worse (C0105122~C0205169~C0332271) prognostic significance associated with more severe ECG abnormalities (C0000768~C0000769) .

In agreement with other reports (C0178935~C0335038) , patients (C0030705) who smoke cigarettes showed a better in-hospital course (C0489547) .
In another study by Barbash et al , more advanced coronary artery disease (C0010068) explained the worse prognosis (C0278252) of nonsmokers .
One explanation might be that smokers (C0037366~C0037369~C0337664) have less-advanced atherosclerotic disease (C0012634~C0028432~C0263648~C0274294) at the time (C0040223~C0183941~C0332311~C0392761~C0449243) of their first clinical episode (C0277793~C0332189) of ischemic disease (C0012634~C0028432~C0263648~C0274294) .

Some common clinical variables (C0439828) , such as diabetes (C0011847) and heart failure (C0018801) , did not appear as predictors in this analysis (C0002778) .
Also , these variables (C0439828) did not turn out to be independent (C0332291) predictors in an analysis (C0002778) involving 3171 patients (C0030705) with non-ST-segment elevation (C0439775) ischemic syndrome (C0039082) in the ESSENCE (C0028910) trial .

We surveyed very evident heart failure (C0018801) with overt respiratory distress (C0476273) only .
Thus the low frequency (C0205213) of heart failure (C0018801) at the time (C0040223~C0183941~C0332311~C0392761~C0449243) of presentation (C0449450) may be a reflection of these diagnostic criteria (C0243161) .
It was not possible to assess either the predictive value (C0042295~C0401806) of exertional unstable angina (C0002965) or that of subacute angina (C0002962) at rest (C0035253~C0175724~C0332153~C0332575~C0439117~C0439142) ( patients (C0030705) without angina (C0002962) within the preceding 48 hours (C0439227) ) because 96.7 % of our patients (C0030705) belonged to Braunwald 's type III (C0441731) classification .

In - Hospital (C0019994) Prognostic Assessment

Recently reported comparisons of early (C0205085) conservative versus early invasive therapeutic (C0039796~C0302350) strategy indicates that there is no benefit associated with the routine use (C0042153~C0439224~C0449865) of any of these strategies .
Therefore it is still necessary to adjust the therapeutic recommendation (C0034866) to the individual needs (C0027552~C0028158~C0332125~C0347979~C0439138~C0439225~C0441922) of each patient (C0030705) .
Because of the lack (C0332268) of other adequate markers (C0181733~C0522501) of instability , clinical risk (C0035647) stratification performed at admission (C0457453) and during hospitalization (C0019993) is still essential to identify patients (C0030705) at high risk (C0332167) in which an early (C0205085) interventional strategy may be recommended .
The findings (C0243095~C0263541~C0332285) of this study are consistent with this appreciation : the appearance (C0233426) of refractory angina (C0002962) , the clinical (C0205210) hallmark of instability , was followed by a rise (C0034601~C0331968) in the rate (C0034693) of interventions and of major events (C0441471) of 44.2 % and 22.7 % , respectively , whereas persistent (C0205322) relief of symptoms (C0178311~C0426409) after admission (C0457453) was associated with a favorable in-hospital (C0019994) outcome and a lower procedural rate (C0034693) ( 11.1 % ) .

Although thrombotic complications (C0009566~C0009567~C0392327) related to procedures (C0025664~C0184661) driven by symptoms (C0178311~C0426409) can be considered to belong to the unstable (C0443343) state , interventions can confound the predictive analyses (C0002778) because recurrence (C0034897) of angina (C0002962) prompts the physician (C0031831) to intervene .
The existence of an intrinsic (C0205102~C0439674) adverse significance associated with refractory angina (C0002962) is supported by our finding (C0243095~C0263541~C0332285) of a significant relation between refractory angina (C0002962) and major (C0205164) nonprocedural ( " spontaneous " ) events (C0441471) .

Limitations (C0449295)

Because no valid universal (C0175671) definition of refractory angina (C0002962) exists , the criteria (C0243161) used in this study may be different from those used by others (C0205394~C0220886~C0237094~C0237098~C0237111~C0237120~C0237127~C0237133~C0237142~C0237155~C0237161~C0237162~C0237169~C0237173~C0237178~C0237185~C0237190~C0237195~C0237206~C0237212~C0237221~C0237228~C0237237~C0237242~C0237243~C0237247~C0237255~C0237262~C0237266~C0237276~C0237283~C0237285~C0237289~C0237300~C0237305~C0237318~C0237324~C0237330~C0237338~C0237343~C0237344~C0237353~C0237357~C0237361~C0237366~C0237369~C0237373~C0237380~C0237387~C0237394~C0237395~C0449210) .
Because of the pragmatic character (C0007952~C0486537) of the study , neither the presence (C0392148~C0392743) of ECG evidence (C0332120) of ischemia (C0022116) nor the use (C0042153~C0439224~C0449865) of heparin (C0019134) during the episodes (C0277793~C0332189) of chest pain (C0008031) were required for the diagnosis (C0011900~C0011901) of refractory angina (C0002962) .

In this study the use (C0042153~C0439224~C0449865) of intravenous heparin (C0354566) was less than what might be anticipated ( 57 % ) because in recent years (C0439234~C0439508) there has been a greater trend (C0040833) toward increased use (C0042153~C0439224~C0449865) .
However , the use (C0042153~C0439224~C0449865) of heparin (C0019134) was higher in patients (C0030705) without precipitating factors ( 69.7 % ) .
This finding (C0243095~C0263541~C0332285) is not unexpected because in primary unstable angina (C0002965) ( Braunwald class B (C0004914~C0004916~C0004922~C0004923~C0005139~C0006030~C0053353~C0205170~C0332272~C0392746~C0439105~C0439127) ) the presence (C0392148~C0392743) of a thrombus (C0087086) can be presumed and the use (C0042153~C0439224~C0449865) of heparin treatment (C0039798~C0087111) is well substantiated .
The studies by Theroux and Wallentin showed the effectiveness (C0205414) of intravenous heparin (C0354566) and low-molecular-weight heparin (C0019139) in preventing major complications (C0009566~C0009567~C0392327) during the acute phase (C0439557) of unstable angina (C0002965) .
Consequently , in our study , a lower incidence (C0021149~C0220856) of refractory angina (C0002962) and a better (C0004914~C0004916~C0004922~C0004923~C0005139~C0006030~C0053353~C0205170~C0332272~C0392746~C0439105~C0439127) outcome could have been observed if heparin (C0019134) had been used more frequently or if a more (C0205172) stringent definition of refractory angina (C0002962) had been used .
Nonetheless , mortality rate (C0034693) in this series (C0205549) was similar to that reported in recent (C0332185) trials .
At 30 days (C0439228~C0439229~C0439505) the death rate (C0205848~C0220817) was 3.6 % and 3.9 % in the heparin arms (C0003792) of the studies ESSENCE (C0028910) and GUSTO II B (C0004914~C0004916~C0004922~C0004923~C0005139~C0006030~C0053353~C0205170~C0332272~C0392746~C0439105~C0439127) , respectively .
In the PURSUIT trial , 30-day mortality rate (C0034693) was 3.7 % in the control arm (C0003792) .
In the TIMI-IIIB study , the death rate (C0205848~C0220817) of the early conservative arm (C0003792) was 4.7 % at 42 days (C0439228~C0439229~C0439505) .
In the PRISM PLUS (C0332287) study , the 30-day mortality rate (C0034693) was 4.0 % in patients (C0030705) treated with heparin (C0019134) .

Clinical (C0205210) Implications

The relative (C0080103~C0205345) benefits of early invasive and early conservative treatment (C0039798~C0087111) strategies have been debated for many years (C0439234~C0439508) .
Either of these approaches (C0449445) is considered clinically appropriate for many patients (C0030705) .
Accordingly , the identification (C0020792) of candidates for a more aggressive (C0001807) strategy has become of utmost importance .
Because of its simplicity and wide availability (C0470187) , clinical and ECG markers (C0181733~C0522501) are widely used for this purpose (C0449256) .
Unfortunately , the specificity (C0037791) of the clinical (C0205210) predictors is insufficient to fulfill the requirements of risk (C0035647) stratification under the current therapeutic modalities (C0278616) .

In recent years (C0439234~C0439508) , serum (C0229671) markers (C0181733~C0522501) such as troponin T (C0077404) and C-reactive protein (C0006560~C0201657) have become the focus (C0205234~C0344236) of interest (C0123730~C0205103) for risk (C0035647) stratification of unstable angina (C0002965) , with promising results (C0332294) .
It is very likely that routine measurements (C0242485) of these markers (C0181733~C0522501) will contribute to the identification (C0020792) of those patients (C0030705) at highest risk (C0332167) for cardiac events (C0441471) .
A major question (C0332147) to be explored is how much information these new biochemical markers (C0206015) will add to what is already provided by the classic clinical (C0205210) predictors .

Acknowledgements

We thank Gianni Tognoni , MD , for his assistance (C0018896) in the design of the study ; Dr Pierre Theroux for helpful comments (C0282411) on early (C0205085) drafts of this paper (C0030351~C0428642) ; and Maria (C0324143) Elena Aizpurua for secretarial assistance (C0018896) .

REFERENCES