Brachial pulse pressure in heart failure: simple to measure but complex to interpret

This editorial refers to €˜Differing prognostic value of pulse pressure in patients with heart failure with reduced or preserved ejection fraction. Results from the MAGGIC individual patient meta-analysis€™ by C.E. Jackson et al., doi:10.1093/eurheartj/ehu490. The paper was inadvertently published in Volume 36, Issue 18 without this editorial. Increased arterial stiffness is expected to affect cardiac work unfavourably and to have a detrimental effect on prognosis, especially in patients with heart failure (HF). Brachial pulse pressure (PP) has been used as a marker of arterial elastic properties to assess the effect of vascular dysfunction on target organ damage and thus to predict adverse outcome. 1€“6 Should PP, a crude surrogate marker of arterial stiffness, be proven to have prognostic value in patients with HF, then clinicians would be provided with a valuable, easy-to-measure, low-cost biomarker with an important pathophysiological background. The study by Jackson et al. from the MAGGIC individual patient meta-analysis investigated the prognostic importance of brachial PP in patients with heart failure and reduced (< 50%) ejection fraction (HF-REF) vs. preserved ejection fraction (HF-PEF). 7 The strong point of this study results from the analysis of individual patient data from 22 studies with an impressive number of 27 046 patients included even though only 18.5% of them had HF-PEF. The authors reported that lower PP independently predicted all-cause mortality in HF-REF patients, confirming previous knowledge on the prognostic value of low PP in patients with HF-REF (probably an index of low stroke volume in these patients). 1, 2 In contrast, in HF-PEF patients, the importance of high PP (probably an index of increased arterial stiffness in these patients) is doubted, in contrast to what might have been expected on the basis of pathophysiology and some previous work. 3 In HF-PEF patients, higher PP was associated with crude mortality but lost its predictive value after adjustment for age, gender, hypertension, diabetes, atrial fibrillation, and ischaemic aetiology, all significant predictors of mortality. First, regarding HF-REF patients, the study by Jackson et al. emphasizes the prognostic importance of low PP (< 53 mmHg in the whole population) in HF-REF patients, in agreement with previous studies. 1, 2 However, a clinically useful cut-off value of PP may be difficult to be derived from the current study as the values of the PP quintiles differ among the various study subgroups. More specifically, the first quintile of PP (< 39 mmHg) is predictive of worse outcome in patients with systolic blood pressure (SBP) < 140 mmHg while the fourth quintile is borderline predictive of better outcome in patients with SBP >140 mmHg (table 5 of the MAGGIC study). Thus, the predictive value of PP in HF-REF should be interpreted according to the corresponding SBP values. To this end, the authors correctly propose that low PP should not be considered as an index of improved arterial elasticity but rather as an index of excessively low stroke volume in HF-REF, and as such is associated with increased mortality. Indeed, in the VMAC study, in which left ventricular ejection fraction (LVEF) increased with tertiles of PP, there was evidence that a relatively preserved LV systolic function is required to maintain an elevated PP. 4 Similarly, in the study of the MAGGIC investigators under discussion here, higher PP was associated with higher SBP and, more importantly, higher LVEF, suggesting that stroke volume is a major determinant of the PP value. In the same context, the opposite role of brachial PP compared with arterial stiffness in the prediction of outcome in patients with HF-REF has been clarified previously; 2 higher pulse wave velocity (PWV)€”a true marker of arterial stiffness€”has been associated with worse outcomes, while higher PP€”here a marker of LV function and stroke volume and not stiffness€”was associated with better outcomes. Secondly, regarding HF-PEF patients, the role of PP in predicting outcome is more complex. In the crude data, higher PP is associated with increased mortality. However, this association loses its statistical significance in a multivariable model including other MAGGIC predictors of outcome (age, gender, hypertension, diabetes, atrial fibrillation, and ischaemic aetiology). These predictors of mortality are mostly co-morbidities associated with increased arterial stiffness and consequently increased brachial PP. 5 Adjustment for these covariates may have eliminated the prognostic significance of high PP in this cohort, because of an overadjustment, as implied by the authors and also supported by the data of this study itself; most MAGGIC variables (age, gender, hypertension, and diabetes) interact to a great degree with PP (table 2 of the MAGGIC study). Furthermore, using brachial PP as a surrogate for aortic PP (and aortic stiffness) is not accurate, because of the PP amplification between central and peripheral arteries; increased PWV along the stiffened arteries and early arrival of wave reflections lead to augmentation of the central aortic SBP and reduction of the diastolic BP. 3 Further research with validated markers of increased arterial stiffness or abnormal wave reflection with established prognostic value, namely central PP, augmentation index, or PWV (already studied in HF-REF), 2 would be useful to clarify the effects of pulsatile arterial wall properties on outcome in HF-PEF patients on the top of other co-morbidities. Several studies have suggested a better predictive value of central PP or PWV for adverse outcome compared with brachial PP in various patient cohorts including HF-REF and HF-PEF patients. 2, 6, 8, 9 Furthermore, data from the study of Jackson et al.7 suggest that even in HF-PEF patients, low PP is a marker of increased risk. Indeed, in the crude data, a J-curve is evident; the lowest PP (< 45 mmHg) and highest PP (> 79 mmHg) quintiles show higher mortality rates compared with the remaining quintiles in mortality curves analysis (figure2 of the MAGGIC study). A significant J-shaped relationship between PP and mortality has recently been reported in HF-PEF patients in the DIG trial, 10 as also acknowledged by Jackson et al . Whether lower PP could be an index of low stroke volume in some HF-PEF patients may be a matter of future research. Reduced stroke volume (e.g. ˆ¼35 mL/m 2 ) has been previously reported in HF-PEF 11 and has also been recognized as a characteristic feature of paradoxical low-flow, low-gradient aortic stenosis with preserved EF, usually presenting with HF-PEF. 12 Although it may not be safe to draw solid conclusions from subgroup analyses, the authors show that in the subgroup of HF-PEF patients with acute HF, lower PP ( < 45 mmHg) is an independent predictor of mortality. Whether low PP in HF-PEF patients presenting with acute decompensated HF, presumably in a worse clinical state, reflects low stroke volume (similarly to HF-REF) cannot be explored in this study. Interestingly, data from the Acute Heart Failure Global Survey of Standard Treatment (ALARM-HF) show that low SBP (< 100 mmHg) is related to in-hospital mortality in acute decompensated HF-PEF. 13 The relationship of SBP to PP and mortality in HF is an important issue that needs attention in the interpretation of this study. The analysis provided by the authors is probably not sufficient to appreciate the complexity of this interaction. For example, a patient presenting with HF, PP of 30 mmHg, and SBP of 150 mmHg (HF is probably triggered by hypertension) is simply very different from a patient with HF, PP of 30 mmHg, and SBP of 90 mmHg (the patient is almost in shock and the very low PP indicates a very low stroke volume and poor prognosis). In the MAGGIC study, when SBP is taken into account (table 5 of the MAGGIC study), PP (low) appears to have a prognostic value only in patients with HF-REF and a SBP <140 mmHg. Moreover, data provided in the appendix show that both PP and SBP, when used as continuous variables and not as quintiles, are independent predictors of mortality but in an opposite direction, in both HF-REF and HF-PEF populations. Decreasing SBP is associated with increased mortality but so is increasing PP, in agreement with previous findings in post-myocardial infarction patients. 14 All the above emphasize the complex interaction between SBP and PP, indicating the need for measurement of validated indices of arterial stiffness. Increased arterial stiffness (i) increases central aortic SBP and therefore LV afterload and cardiac work; and (ii) reduces diastolic BP leading to reduced coronary perfusion, 3 and may thus contribute to the pathogenesis and exacerbation of HF ( Figure 1 ). Increased arterial stiffness has been associated with exercise intolerance in HF-REF and HF-PEF, 3, 15 and may contribute to poor prognosis in these patients. 2 The need to study arterial stiffness and wave reflections on admission and also pre-discharge to account for the effects of treatment 6 in relation to HF outcome is imperative. In order to do that, appropriate tools should be used to avoid misleading simplifications. Devices using a simple brachial cuff or a tonometry sensor are currently available and can provide both central and peripheral pressures even in the setting of acute dyspnoea. 16 Proposed association between vascular indices and LV function with outcome in patients with heart failure (HF) and reduced or preserved ejection fraction (HF-REF or HF-PEF). BP, blood pressure. In conclusion, PP is known to be a complex biomarker affected by LV function, stroke volume, as well as arterial elastic properties. HF is also a complex syndrome with various phenotypes difficult to discriminate€”particularly HF-PEF. The prognostic usefulness of a single brachial PP value may be evaluated only if the concomitant values of LVEF (or preferably stroke volume) and SBP are known. Exploring the role of PP in HF is important as it may contribute in understanding the pathophysiology of the syndrome, but its role as a clinical biomarker (with clear cut-off limits) may be limited. Whether central aortic pressure, suggested as a powerful and potentially more robust predictor of risk than brachial BP, 8, 9 may be used as a prognostic marker in HF remains to be proven. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.