Preinfarction angina، INTRODUCTION AND OBJECTIVES: Preinfarct angina can reduce the size of myocyte necrosis and improve the prognosis after acute myocardial infarction. The aim of this study is to analyze total mortality 6 months after acute myocardial infarction according to the presence or absence of preinfarct angina.
METHODS:We prospectively included 175 consecutive patients with acute myocardial infarction, 72 (41.1%) with preinfarct angina.
The follow-up was carried out for 6 months. In the group without preinfarct angina, 16 patients died in the follow-up (15.5%) compared to seven (9.7%) in the group with preinfarct angina (log-rank: 1.03, p = 0.311). The risk functions of the two groups show a higher risk of mortality throughout the follow-up in the group without pre-infarct angina. In the multiple logistic regression model, the presence of preinfarct angina does not contribute significantly to reduce mortality (OR = 0.43, 95% CI = 0.09-2.22, p = 0.303).
A significant interaction was detected between the consumption of sulfonylureas prior to infarction and preinfarct angina (p = 0.017).
CONCLUSION:In this study, we did not find significant differences in total mortality after 6 months of follow-up after acute myocardial infarction according to the presence or absence of preinfarct angina. The risk of death, however, appears to be increased in patients without preinfarct angina throughout the follow-up period. There is a significant interaction between the consumption of sulfonylureas prior to infarction and pre-infarct angina.
Several studies have analyzed the impact on mortality of the presence of preinfarct angina (PA) with discordant results. Thus, there are published studies that find no differences in in-hospital mortality when patients are grouped according to the presence or absence of AP 1 , others that find a reduction in in-hospital mortality in patients with AP 2-5 , at least in patients younger 6 and others who find an increase in in-hospital mortality if the patient has AP 7 .
The papers that analyze mortality in the medium / long term are more concordant, observing an accumulated increase in mortality in patients with PA that equals 3.8or it exceeds that of patients without AP 1,2,7,9 .
This paper intends to prospectively analyze the mortality in the medium term (6 months) after an acute myocardial infarction (AMI) in a cohort of patients divided according to the presence or absence of PA. The results of the study complement those already published by our group on the effect of PA on the size of myocyte necrosis 10 .
The methodology of the study has already been described previously 10. Briefly, AMI was diagnosed when the patient met the following three requirements: pain, discomfort or chest tightness attributable to myocardial ischemia of at least 30 min duration; changes in the electrocardiogram suggestive of acute myocardial ischemia (both elevation and depression of the ST segment); elevation in the creatine phosphokinase (CK) and CK-MB levels at least twice above the value considered by our laboratory as normal (195 U / l for CK and greater than 20 U / l for CK-MB).
All the patients were interrogated at the time of admission to determine if they presented unstable angina before the infarction. Pre-infarct angina was defined as the presence of pain, oppression or thoracic discomfort at rest in the 7 days prior to the infarction. Age was analyzed,
6 , the presence of cardiovascular risk factors (hypertension, diabetes mellitus, hypercholesterolemia and tobacco consumption), the presence of previous AMI and previous coronary interventions (angioplasty and revascularization surgery), the pharmacological treatment that the patient consumed in the 7 days prior to infarction,
delay in going to hospital AMI location and administered treatment, infarct size by area under the curve CK-MB in the first 24 h following the method of Kloner 4, the highest degree of Killip reached, the mechanical complications (ventricular septal defect, free wall rupture or acute mitral regurgitation) and electrical complications (supraventricular and ventricular tachyarrhythmias of at least 3 beats and second or third degree atrioventricular blocks).
The ejection fraction (EF) calculated before discharge was categorized into two categories: significant systolic dysfunction (FE ² 0.40) and mild dysfunction or normal function (EF> 0.40). The patients discharged from the hospital were contacted at 3 and 6 months after the infarction, either through a personal interview or through telephone contact. The only final objective was to determine the death of any cause in a follow-up of 6 months,
although the reinfarction, revascularization and re-entry variables were also analyzed for unstable angina. a posteriori in two groups, according to whether or not they had died in the follow-up period.
The quantitative variables with normal distribution are exposed with the mean and between brackets the standard deviation.
If the distribution was not normal, interquartiles 25 and 75 are presented with the median and between parentheses. For the comparison of continuous quantitative variables, parametric tests (Student’s t) were used, unless the variable did not follow a normal distribution,
in which case non-parametric tests (Mann-Whitney U) were used and for qualitative variables the test of c <+ f $> 2with the Welch correction. Fisher’s exact test was applied when necessary. Mortality curves were analyzed using the actuarial curve and its risk function (dividing the follow-up period into one-month intervals) or through the Kaplan-Meier curves. The comparison between curves was carried out using the Mantel-Haenszel test (log-rank tests).
A logistic regression model was constructed with the variable “death” as a dependent and with the variable “preinfarct angina” as independent (the Cox proportional hazards model was not used because the risk of death is not constant throughout the entire period. follow-up) (Fig. 1). If any of the variables did not follow a normal distribution, a logarithmic transformation was performed and it was again verified that their distribution adjusted to a normal curve.
A maximum model was created with the variables of confusion and interaction, evaluating first the latter according to whether their contribution to the model was significant or not, and subsequently the confounding variables. The interaction variables analyzed were those already