The higher, the better? Blood pressure after acute myocardial infarction

It is of pre-eminent importance for the managing clinician to stratify the risk of myocardial infarction patients with respect to future cardiovascular morbidity and, in particular, cardiovascular mortality. The size of a myocardial infarction, the ejection fraction, the degree of myocardial revascularization and the status of the autonomic nervous system (and related arrthythmogenicity) are all established predictors that enable assessment of the individual risk of such patients. Likewise, modifiable risk factors predict future events and there is no doubt that treatment of hyperlipidaemia, stringent glycaemic control in diabetics and abstinence from smoking lead to a reduction in cardiovascular morbidity and mortality [1,2]. Moreover, the prevalence of arterial hypertension is high in myocardial infarction patients compared to the normal population [3]. Therefore, the need for blood pressure lowering by medication is inferred as a measure of secondary prevention in patients with hypertensive [4] or even pre-hypertensive blood pressure levels [5]. Therefore, as reported in this issue of the journal, at first glance, it appears to be paradoxical that Yap et al. [6] correlate high blood pressure values during the phase of hospitalization for myocardial infarction with a decreased medium to long-term cardiovascular mortality. The authors carried out a retrospective analysis on more than 3000 placebo-treated patients of the EMIAT, CAMIAT, SWORD, TRACE and DIAMOND–MI studies with ejection fractions < 40% or documented asymptomatic ventricular arrhythmias [6]. The data suggest that higher systolic and diastolic blood pressure values during hospitalization after myocardial infarction are associated with a decreased all-cause mortality and a raised rhythmogenic mortality. The investigation is interesting with respect to the idea of testing whether a vital parameter measured routinely during the hospitalization stay is predictive for the prognosis. Although such an investigation may be obvious, it was conducted systematically with this extension in patients who were hospitalized for myocardial infarction by stringent retrospective meta-analysis for the first time. By contrast, the association between blood pressure levels and survival has been studied extensively during past decades. In this context, several population-based approaches [7–10], as well as multicenter studies in individuals with hypertension and myocardial infarction [11–14], have demonstrated a J-shaped or U-shaped association between blood pressure levels and future mortality (J-curve phenomenon). Several mechanisms may explain the increase of future risk with lower blood pressure: (i) lower blood pressure may compromise blood flow to target organs impairing particularly coronary perfusion in patients with incomplete revascularization; (ii) low blood pressure may be an epiphenomenon related to underlying chronic illness (i.e. severity of heart failure) and, thus, only indirectly relate to increased morbidity and mortality; and (iii) low diastolic pressure may result from an increase in pulse pressure, reflecting stiffness of large arteries and representing a marker for advanced vascular disease. Taken together, these and other mechanisms explain why mortality may increase with lower blood pressure, a finding that is supported by the work of Yap et al. [6]. On the other hand, there is no doubt that chronically-elevated blood pressure in patients with myocardial infarction also increases mortality. Several approaches consistently have demonstrated an increase in morbidity and mortality with higher blood pressure in patients with coronary artery disease [3,7,8]. Such findings apparently oppose the data of Yap et al. [6]. The current work thus requires an explanation with respect to the apparent inverse association between outcome and blood pressure values. First, and most importantly, from their analysis, it is unclear whether the association observed by Yap et al. [6] was truly linear (i.e. the higher, the better) or curvilinear (i.e. J- or U-shaped). In other words, the data do not reveal whether the risk decrease of approximately 15–20% for every 10 mmHg increase in blood pressure was evident in the entire spectrum of blood pressure variability. Therefore, a simple explanation for the paradox may be a strong negative effect on survival from patients with large myocardial infarctions, severe heart failure and thus very low blood pressure. In a linear correlation, this effect may mask a plateau or even a decrease of risk at optimal blood pressure levels. Thus, it remains unclear whether there is a direct correlation between the higher systolic and diastolic blood pressure values and the decreased mortality of the myocardial infarction patients, or whether the correlation is dominated by strong deviation in patients with reduced left ventricular function. Second, it is also conceivable that the low blood pressure leads to a reduced dosage of angiotensin-converting enzyme inhibitors, β-blockers, or other prognostically relevant medications ultimately leading to a deterioration of the prognosis. Third, all clinical studies contributing to this meta-analysis were carried out in the era of thrombolytic therapy of patients with myocardial infarction. Coronary angiography was only performed in some patients. It is thus unclear whether the results are also valid for myocardial infarction patients undergoing complete interventional revascularization. Particularly, Messerli et al. [14] demonstrated that low blood pressure is associated with poor prognosis only in those patients with incomplete restoration of coronary perfusion. Fourth, the study by Yap et al. [6] focused on a subgroup of myocardial infarction patients with a relatively high risk: patients with reduced left ventricular function and patients with a major arrhythmogenic potential. For this reason, the results presented may not be applicable for clinical purposes to all myocardial infarction patients. Finally, many of the limitations in the interpretability of the results are due to the retrospective design of the present meta-analysis and the inhomogeneities of patient populations that are unavoidable in such studies. Apart from theoretical considerations that govern risk stratification of myocardial infarction patients in clinical practice, clinicians are often far from being precise with respect to their prediction in individual patients. For this reason, all possible avenues to improve risk prediction must be explored to improve patient prognosis after myocardial infarction. In this respect, the study reported by Yap et al. [6] shows an interesting phenomenon that must be confirmed in future investigations. We certainly learn that low blood pressure in the first days after myocardial infarction is bad news for such patients. Further details of this concept must be developed and refined with respect to the underlying mechanisms, as well as the clinical significance of such findings with respect to the global risk stratification of myocardial infarction patients.