Organ perfusion pressure at admission and clinical outcomes in patients hospitalized for acute heart failure

Abstract

Aims

Hypoperfusion portends adverse outcomes in acute heart failure (AHF). The gradient between end-organ inflow and outflow pressures may more closely reflect hypoperfusion than mean arterial pressure (MAP) alone. The aim of this study was to investigate organ perfusion pressure (OPP), calculated as MAP minus central venous pressure (CVP), as a prognostic marker in AHF.

Methods and results

The Sodium NItroPrusside Treatment in Acute Heart Failure (SNIP)-AHF study was a multicentre retrospective cohort study of 200 consecutive patients hospitalized for AHF treated with sodium nitroprusside. Only patients with both MAP and invasive CVP data available from the SNIP-AHF cohort were included in this analysis. The primary endpoint was to assess OPP as a predictor of worsening heart failure (WHF), defined as the worsening of signs and symptoms of heart failure leading to intensification of therapy at 48 h. One hundred and forty-six patients fulfilling the inclusion criteria were included [mean age: 61.1 ± 13.5 years, 32 (21.9%) females; mean body mass index: 26.2 ± 11.7 kg/m²; mean left ventricular ejection fraction: 23.8%±11.4%, mean MAP: 80.2 ± 13.2 mmHg, and mean CVP: 14.0 ± 6.1 mmHg]. WHF occurred in 14 (9.6%) patients. At multivariable models including hemodynamic variables (OPP, shock index, and CVP), OPP at admission was the best predictor of WHF at 48 h [OR 0.91 (95% confidence interval 0.86–0.96), P-value = 0.001] with an optimal cut-off value of 67.5 mmHg (specificity 47.3%, sensitivity 100%, and AUC 0.784 ± 0.054). In multivariable models, including univariable significant parameters available at first bedside assessment, namely New York Heart Association functional class, OPP, shock index, CVP, and left ventricular end-diastolic diameter, OPP consistently and significantly predicted WHF at 48 h.

Conclusion

In this retrospective analysis on patients hospitalized for AHF treated with sodium nitroprusside, on-admission OPP significantly predicted WHF at 48 h with high sensitivity.

Structured Graphical Abstract

Organ perfusion pressure (OPP) in acute heart failure.

Accuracy of OPP, mean arterial pressure (MAP), shock index, and central venous pressure to predict worsening heart failure at 48 h. *OPP and MAP are inversely associated with the primary outcome.

Open in new tabDownload slide

Introduction

Acute heart failure (AHF) is a growing health a dismal prognosis, yielding a mortality rate of 26.7% at 1 year.¹ AHF is typically characterized by signs and/or symptoms of pulmonary congestion due to increased cardiac filling pressures but may also be accompanied by venous fluid collection due to sodium and water retention and/or right ventricular failure.² The presence of hypoperfusion alongside systemic congestion yields an even worse prognosis with increased mortality.^3,4 In fact, the complex interplay between low arterial and high central venous pressure (CVP) in AHF can influence organ function adding backward congestion to a low anterograde perfusion status, leading to renal, hepatic, and intestinal failure, especially in advanced decompensated heart failure, which is characterized by persistent signs and symptoms of HF despite maximal therapy.⁵ Worsening renal function (WRF) is a strong independent predictor of adverse outcomes in advanced decompensated HF and more closely reflects a state of increased venous congestion rather than impaired cardiac output, contributing to worsening HF (WHF) recurrences.² Indeed, episodes of WHF, defined as the aggravation of symptoms and signs of HF in patients with pre-existing HF, requiring intensification of treatment, characterize the clinical course of HF, and portend further increased mortality.⁶ The abdominal compartment might contribute significantly to deranged cardiac and renal function in congestive HF; increased intra-abdominal pressure with reduced abdominal perfusion pressure due to abdominal congestion predicts adverse outcomes due to multiorgan dysfunction, including reduced renal perfusion pressure, impaired liver function, and intestinal barrier damage.^7–10 Also, situations that increase the intrathoracic pressure, such as positive end-expiration pressure, must also be considered in this setting as they may pose an increased pressure load on the venous congestive state.¹¹ Ultrasound assessment of the volume status with the multiparametric VExUS approach may come as a major support in carefully evaluating the venous congestive state of HF patients.¹²

Organ perfusion pressure (OPP), calculated as the difference between mean arterial pressure (MAP) and CVP, was shown to provide prognostic value in different clinical situations,^13–15 as it may serve as a simple and ‘all-inclusive’ index to identify patients with refractory congestion and a major risk of end-organ damage. Thus, the aim of the present study was to investigate OPP as a novel risk marker in patients hospitalized for AHF.

Methods

This study was conducted on the Sodium NItroPrusside Treatment in Acute Heart Failure (SNIP-AHF) study population. Details on the SNIP-AHF study have been previously published elsewhere (www.ClinicalTrials.gov registration number: NCT05027360).¹⁶ Briefly, SNIP-AHF was a retrospective multicentre study including all consecutive adult patients admitted in a Cardiac Intensive Care Unit for AHF with reduced ejection fraction requiring intravenous sodium nitroprusside between January 2016 and January 2020 at two Italian centres (ASST Niguarda in Milan and A.O.U. Città della Salute e della Scienza in Turin). Patients with post-cardiotomy AHF and those with cardiogenic shock requiring high dosage vasoactive agents (defined as a vasoactive-inotropic score > 10¹⁷) and/or short-term extracorporeal life support at the time of sodium nitroprusside initiation were excluded.

Patients’ demographics (age and sex), clinical, hemodynamic and laboratory values, echocardiographic parameters, adverse events, and clinical outcomes were obtained by chart review. Laboratory, hemodynamic, and echocardiographic data were recorded before intravenous sodium nitroprusside initiation (T0), 48 h later (T1), and at discharge.

For the sake of the present analysis, only patients with both MAP and CVP data available were included. Blood pressure measurements were derived either non-invasively or invasively when an arterial line was in place; the latter was preferred when available. CVP was derived from the recordings of a central venous catheter with radiologically confirmed correct positioning at the junction of the superior cava vein with the right atrium. MAP and CVP were measured simultaneously. OPP was calculated as MAP minus CVP.

The primary endpoint of this study was to assess OPP as a predictor of worsening heart failure (WHF) at T1 as compared with MAP, shock index, and CVP alone. WHF was defined as the worsening of signs and symptoms of heart failure (refractory pulmonary or systemic congestion, and/or hypotension with organ hypoperfusion), as adjudicated by the caring physician, requiring initiation or intensification of inotropic agents and/or up-titration of the diuretic therapy started on admission.^6,18 Shock index was defined as the ratio of heart rate to systolic blood pressure. Secondary endpoints included WRF at T1 (defined as an elevation in serum creatinine levels ≥ 0.3 mg/dL²), persistent hepatic damage at T1 (defined as high total bilirubin at T1), significant N-terminal pro-B-type natriuretic peptide (NT-proBNP) reduction at T1 (defined as a reduction of at least 25% compared with baseline), a composite endpoint of WHF or non-significant NT-proBNP reduction at T1 (defined as a reduction of lesser than 25% compared with baseline), and in-hospital all-cause death. High total bilirubin was defined as serum total bilirubin ≥ 1.2 mg/dL, as values above this threshold were found to predict adverse outcomes in advanced HF patients.^19,20 A composite renal outcome consisting of an elevation in serum creatinine levels ≥ 0.3 mg/dL or a reduction of urinary output < 0.5 mL/kg/min at T1 was also assessed; also, an end-organ function worsening outcome was defined as an increase in Sequential Organ Failure Assessment (SOFA) score at T1 as compared with T0. The role of pulse pressure (systolic blood pressure minus diastolic blood pressure) and coronary perfusion pressure (diastolic blood pressure minus CVP) on WHF was also assessed.

The study was conducted in accordance with ethical principles based on Helsinki’s Declaration, International Conference on Harmonization for Good Clinical Practice, and the current ethical rules. This study was approved by the Local Ethics Committee of Milano Area 3 of the ASST Grande Ospedale Metropolitano Niguarda (Milan, Italy) (Approval number: S_04_06_21_5186). Due to the retrospective nature of the study, the requirement for informed consent was waived by the Institutional Review Board.

Statistical analysis

Continuous variables were reported as mean (standard deviation) or median (interquartile range), as appropriate, while categorical variables were reported as percentages. The presence of normal distribution was verified by the Shapiro–Wilk test.

A multivariable logistic backward conditional model was built to test OPP against other hemodynamic parameters available at initial bedside assessment, namely shock index, MAP, and CVP, as the best predictor of WHF at T1. Receiver operating characteristic (ROC) curves were elaborated, and the associated areas under the curves (AUCs) were calculated and compared with determine which clinical parameter was best associated with the primary endpoint. An optimal cut-off value for OPP was then elaborated and defined as the value with the highest specificity and sensitivity to detect WHF by means of Youden’s index. Multicollinearity among the variables included in the final multivariable model was assessed by calculating the variance inflation factors. When significant collinearity was detected (i.e. variance inflating factor >5), only the independent variable with the highest AUC was retained in the model (see Supplementary material online, Table S1).

Univariable analysis to assess predictors of WHF at T1 was also performed. OPP was then tested against in-study predictors of WHF at T1 at multivariable logistic regression analysis. Variables with a univariable P-value < 0.05 and that were immediately available at first basal bed-side assessment were entered into the model. Sensitivity analyses were also performed including laboratory data in the models. Also, different multivariable models were constructed a priori using combinations of clinically meaningful baseline features (OPP, NT-proBNP, creatinine, left ventricular ejection fraction, and tricuspid annular plane systolic excursion) rather than backwards selection. Results are reported as odds ratios (ORs) with 95% confidence intervals (CIs). Patients with high OPP were then compared against those with low OPP to assess differences between these study subgroups; for the purpose of this analysis, OPP was classified as high or low according to its value being higher or lower than the median OPP value of the study population. A sensitivity analysis was also conducted classifying OPP as either high or low according to its optimal threshold value found at ROC analysis. A separate analysis assessing the incidence of WHF, WRF, and persistent hepatic damage according to OPP quintile distribution was performed. A two-sided P-value < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS 26.0 (SPSS, Chicago, IL, USA).

Results

Among the 200 patients included in the SNIP-AHF study, 54 individuals did not have baseline CVP or MAP data available and were excluded; a total of 146 patients were included in the present study. Detailed characteristics of the study population are reported in Table 1. Overall, mean age was 61.3 ± 13.5 years; 32 (22%) patients were female; mean body mass index was 26.2 ± 11.7 kg/m²; and 128 (88%) patients had advanced decompensated heart failure. On admission, mean MAP was 80.2 ± 13.2 mmHg, mean heart rate was 86.2 ± 18.2 b.p.m., and mean CVP was 14.0 ± 6.1 mmHg. Median OPP was 64.0 mmHg (interquartile range 14.0 mmHg). On echocardiography, mean left ventricular ejection fraction was 23.8% ± 11.4%, mean tricuspid annular plane systolic excursion was 14.9 ± 3.8 mm, mean right ventricular fractional area change was 25.3% ± 6.7%, and mean estimated pulmonary artery systolic pressure was 52.5 ± 16.1 mmHg. The temporal trends of hemodynamic and laboratory data throughout the hospitalization period are depicted in Supplementary material online, Figures S1 and S2.

Primary outcome

The primary outcome of WHF occurred in 14 (9.6%) patients (Table 2). At the multivariable logistic backward conditional model including hemodynamic variables (OPP, shock index, and CVP), OPP at admission was the best predictor of WHF at T1 with an OR of 0.91 (95% CI 0.86–0.96, P-value = 0.001) (Table 3). MAP was not included in the model due to multicollinearity with OPP (see Supplementary material online, Table S1). At ROC curves analysis, the AUC for predicting the primary outcome with OPP was 0.784 ± 0.054 (95% CI 0.679–0.890), which was numerically but not significantly greater than that of MAP [AUC: 0.735 ± 0.068 (95% CI 0.602–0.869), P-value for comparison = 0.540], shock index [AUC: 0.666 ± 0.076 (95% CI 0.518–0.814), P-value for comparison = 0.186], and CVP [AUC: 0.608 ± 0.079 (95% CI 0.453–0.763), P-value for comparison = 0.058] (Structured Graphical Abstract). The OPP cut-off value which provided the best accuracy for predicting WHF was 67.5 mmHg, with a specificity of 47.3% and a sensitivity of 100%.

At univariable analysis, compared with patients without WHF, those experiencing WHF had worse baseline New York Heart Association (NYHA) class (3.7 ± 0.7 vs. 3.2 ± 0.6, P-value = 0.009), lower MAP (70.7 ± 10.8 vs. 81.4 ± 13.1 mmHg, P-value = 0.004), OPP (53.9 ± 9.2 vs. 67.7 ± 14.6 mmHg, P-value = 0.001), higher CVP (16.8 ± 6.7 vs. 13.7 ± 5.9 mmHg, P-value = 0.072), and shock index (0.91 ± 0.19 vs. 0.80 ± 0.19 b.p.m./mmHg, P-value = 0.031) (Table 4); also, patients with WHF had lower serum sodium (P-value = 0.023) and higher values of NT-proBNP (P-value = 0.042). At the multivariable logistic regression models on the primary outcome including clinical data immediately available at first bedside assessment, namely NYHA functional class, OPP, shock index, CVP, and left ventricular end-diastolic diameter, OPP consistently and significantly predicted WHF at T1 across all models (Table 5). In both patients with and without WHF, there seems to be a progressive slight increase in OPP values throughout the index hospitalization, although patients without WHF have significantly greater OPP values at all time points compared with those experiencing WHF (see Supplementary material online, Figure S1).

Sensitivity analyses on the primary outcome

Based on the results of the univariable analysis, multivariable logistic regression models on the primary outcome were built, including NYHA class, OPP, serum sodium, high bilirubin, and NT-proBNP (see Supplementary material online, Table S2). In all backward-selection multivariable models, OPP significantly predicted absence of WHF at T1 with an OR ranging from 0.91 to 0.92.

According to the multivariable models built upon clinical knowledge, OPP significantly predicted absence of WHF at T1 with an OR ranging from 0.89 to 0.91 (see Supplementary material online, Table S3).

Secondary outcomes

WRF at T1 occurred in 14 (9.6%) individuals. Central venous oxygen saturation was the only univariable predictor of WRF (P-value = 0.091) (see Supplementary material online, Table S4). Also, OPP did not predict the composite renal outcome at T1 at univariable analysis (see Supplementary material online, Table S5).

Out of 117 patients with available data, persistent hepatic damage at T1 occurred in 67 (57.2%) patients. At univariable analysis, variables significantly associated with persistent hepatic damage at T1 included the presence at baseline of worse NYHA functional class (P-value = 0.005), lower MAP (P-value = 0.008), OPP (P-value = 0.001), central venous oxygen saturation (P-value = 0.011), higher CVP (P-value = 0.036), baseline total bilirubin (P-value < 0.001), and left ventricular end-diastolic diameter (P-value = 0.001) (see Supplementary material online, Table S6).

Ninety-seven (66.9%) patients showed significant NT-proBNP reduction at T1, which was significantly associated with the baseline presence of renin–angiotensin–aldosterone system inhibitors (P-value = 0.014) and an implantable cardioverter-defibrillator (P-value = 0.039), higher haemoglobin levels (P-value = 0.046) and greater left ventricular end-diastolic diameter (P-value = 0.026) at univariable analysis (see Supplementary material online, Table S7). Fifty-four (37.0%) patients experienced the composite endpoint of WHF or non-significant NT-proBNP reduction at T1; univariable predictors of this composite outcome were still the baseline presence of renin–angiotensin–aldosterone system inhibitors (P-value = 0.016) and an implantable cardioverter-defibrillator (P-value = 0.049) (see Supplementary material online, Table S8). OPP was not significantly different between patients experiencing overall end-organ function worsening (defined by an increase in SOFA score) and those who did not (64 ± 10 vs. 67 ± 16 mmHg, P-value 0.254).

In-hospital death occurred in six (41.7%) individuals. Significant univariable predictors of in-hospital death were lower OPP (P-value = 0.052), central venous oxygen saturation (P-value < 0.001), higher CVP (P-value = 0.009), and creatinine levels (P-value = 0.018) (see Supplementary material online, Table S9).

High OPP vs. low OPP

OPP was defined as either high or low according to its value being higher or lower than the on-admission OPP median value of 64.0 mmHg. Patients with high OPP, as compared with those with low OPP, had lesser prevalence of advanced decompensated heart failure (81% vs. 95%, P-value = 0.012) and atrial fibrillation (38% vs. 59%, P-value = 0.010), lower baseline SOFA score (3.3 ± 2.2 vs. 5.2 ± 2.2, P-value < 0.001), greater serum sodium concentration (138 ± 5 vs. 134 ± 6 mmol/L, P-value < 0.001), lower NT-proBNP [4981 (IQR 6193) vs. 8512 (IQR 12 200) pg/mL, P-value = 0.002], and blood urea nitrogen (72 ± 48 vs. 96 ± 61, P-value = 0.025) levels (Table 1). Moreover, patients with low OPP received significantly higher doses of intravenous furosemide compared with those with high OPP (203 ± 151 vs. 143 ± 134, P-value = 0.013). Twelve (16%) patients with low OPP experienced WHF as compared with two (3%) of those with high OPP (P-value = 0.009) (Table 2). Patients with low OPP had significantly higher incidence of persistent hepatic damage (69% vs. 45%, P-value = 0.007) at T1 compared with individuals with high OPP, whereas no significant difference was found regarding the other secondary outcomes.

When the threshold value of OPP detected at ROC curve analysis was considered, patients with OPP lower than 67.5 mmHg had significantly greater incidence of WHF (17% vs. 0%, P-value = 0.001), persistent hepatic damage (69% vs. 41%, P-value = 0.003), WHF or non-significant NT-proBNP reduction (46% vs. 28%, P-value = 0.030), and in-hospital all-cause death (7% vs. 0%, P-value = 0.038) (see Supplementary material online, Table S10). When WHF was evaluated according to the quintile distribution of OPP, a progressive decrease of WHF incidence was found at increasing quintiles, with no WHF episodes in the two highest quintiles (see Supplementary material online, Figure S3). No specific trend was observed regarding WRF and persistent hepatic damage.

Pulse pressure and coronary perfusion pressure

On univariable analysis, coronary perfusion pressure was significantly lower in patients with WHF as compared with those without (41.3 ± 8.3 vs. 53.6 ± 13.7 mmHg, P-value = 0.001), while no significant difference regarding pulse pressure was found (see Supplementary material online, Table S11).

Discussion

The main findings of this retrospective study can be summarized as follows: among 146 patients hospitalized for AHF with reduced ejection fraction requiring intravenous sodium nitroprusside, OPP on admission significantly predicted WHF at 48 h at multivariable analysis with an AUC of 0.784 ± 0.054, outperforming that of MAP, shock index, and CVP; the OPP cut-off value with the best accuracy for predicting WHF was 67.5 mmHg, with a specificity of 47.3% and a sensitivity of 100%; patients with low OPP had significantly higher incidence of persistent hepatic damage at 48 h compared with individuals with high OPP at univariable analysis.

AHF is a complex syndrome characterized by variable clinical profiles and different underlying pathophysiological processes. Low systolic blood pressure on admission has long been demonstrated to yield poor prognosis despite medical therapy.²¹ Indeed, patients presenting with arterial hypotension may be more likely to carry an advanced disease with low cardiac output and signs of organ hypoperfusion, especially when other ‘cold’ profile signs coexist.⁴ The presence of venous congestion was also proved to be a strong marker of adverse outcomes in AHF.² A retrospective analysis of the SOLVD trial demonstrated that the presence of elevated jugular venous pressure at clinical examination was significantly associated with an increased risk of AHF hospitalizations and death from pump failure in a vast cohort of AHF patients.²² At organ level, capillary blood flow depends on the gradient between MAP and CVP; although under normal conditions autoregulatory mechanisms allow to maintain a stable MAP despite highly variable cardiac outputs, CVP increases can significantly affect tissue blood flow, particularly during low MAP conditions.²³ This ‘double-hit phenomenon’, portraying the complex interplay between low arterial and high venous pressure in AHF patients, can influence organ function adding backward congestion to a low anterograde perfusion status, leading to renal, hepatic, and intestinal failure. This translates into a catastrophic response with increase in congestive status, worsening of HF symptoms and systemic inflammatory response with additional negative effects on cardiac and end-organ function.

OPP is a simple and reproducible parameter resuming the complex interaction between anterograde hypoperfusion and backward congestion; its measurement requires the determination of MAP from either an arterial line or a non-invasive blood pressure cuff and CVP from a central venous catheter or potentially from non-invasive echocardiographic measurements. Its prognostic yield has already been described regarding acute kidney injury among cohorts of critical care or cardiac surgery patients.^14,15 However, to our knowledge, this is the first proof-of-concept study to demonstrate that lower levels of OPP are significantly associated with in-hospital WHF in patients admitted for AHF, proving even superior to both shock index and CVP at multivariable analysis. In-hospital WHF represents a consistent risk factor for hospital readmissions and in-hospital mortality; indeed, a sub-analysis of the ADHERE registry revealed that the occurrence of in-hospital WHF was associated with a greater than two-fold mortality rate at 30 days, a 24% greater risk of all-cause hospital readmissions at 30 days and higher hospitalization costs.²⁴ Due to its high sensitivity, OPP on admission may help identify those individuals at risk of worsening congestion, for whom more strict surveillance with aggressive treatment goals should be established. Interestingly, on-admission OPP outperformed commonly acknowledged predictors of adverse outcomes in AHF, such as MAP and CVP. This might be because neither MAP nor CVP alone, but rather the interaction of the two, i.e. OPP, can truly embrace the ‘double-hit phenomenon’ which affects organ perfusion in AHF patients.²⁵ Thus, in patients with AHF who will be treated with nitroprusside, lower OPP is associated with a greater chance of WHF. The combination of OPP with other parameters of clinical and echographic assessment (echocardiography, lung echography, and VExUS) might provide a valuable tool for better risk stratification and patient management. If low OPP is due to exceedingly high CVP with preserved MAP, diuretic therapy with fluid restriction, and venodilation may be the cornerstone of treatment; conversely, should low OPP depend upon low MAP with normal/low CVP in the setting of AHF, inotropic therapy may rather be needed to treat an underlying condition of hypoperfusion. Nevertheless, OPP alone cannot guide patient management, but a comprehensive multiparametric approach is warranted to define a correct diagnostic and therapeutic strategy.

Conditions raising intrathoracic pressure, such as non-invasive ventilation, obesity, diminished chest compliance (i.e. post-cardiac surgery), and abdominal hypertension, might negatively affect OPP by posing an increased afterload to the venous return, thus increasing the CVP and potentially reducing MAP.¹⁸ Extravasation of fluids in the abdominal interstitium reduces the abdominal wall compliance, further increasing the intra-abdominal pressure, impairing lung distension, and recruitment and decreasing kidney perfusion with greater risk of WRF, altogether leading to WHF and vicious hemodynamic decline. In the present study, however, no significant difference was found regarding body mass index or the prevalence of prior coronary artery bypass surgery or non-invasive ventilation amongst patients with low vs. high OPP, perhaps due to the limited sample size.

Noteworthily, low OPP on admission was associated with persistent hepatic damage at 48 h. An analysis of the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study with Tolvaptan (EVEREST) cohort,²⁶ including patients hospitalized for AHF with reduced ejection fraction, identified sustained total bilirubin elevation as a powerful predictor of all-cause mortality and combined cardiovascular mortality or first AHF hospitalization. Also, persistent hepatic damage was shown to predict death from pump failure in a previous study by Wu et al.¹⁹ stressing the key role of liver function evaluation in AHF and most importantly in more compromised and advanced scenarios. Nevertheless, the low number of recorded events and the observational nature of the study make these findings speculative and hypothesis-generating rather than conclusive.

Interestingly, kidney and liver functions did not significantly differ between patients with low OPP and those with high OPP at baseline, except for significantly higher values of blood urea nitrogen and numerically greater levels of creatinine and total bilirubin in the ‘low OPP’ group; moreover, and perhaps unexpectedly, OPP on admission was not associated with WRF at 48 h; however, these findings might be explained by the small sample size and the low event rate of the present study.

Limitations

This study has some limitations. It is an observational retrospective study on a cohort of highly selected patients from two advanced heart failure units and its results may not be extended to an ‘all-comers’ AHF population; indeed, this cohort represented a more advanced heart failure stage, as 88% of patients were admitted for advanced decompensated heart failure, mean left ventricular ejection fraction was 23.8% ± 11.4%, and mean right ventricular fractional area change was 25.3% ± 6.7%. All patients were treated with the intravenous sodium nitroprusside, which is not currently used in all heart failure units; also, as the peak dose of sodium nitroprusside achievable is often limited by systemic hypotension, one would expect those with higher systemic blood pressure to be able to achieve higher doses of therapeutic nitroprusside, narrowing generalizability of the data. Sample size was small, thus limiting the statistical inferential power of the analyses; moreover, event-rate was low and massive amounts of variables have been compared with that some associations may be spurious; as such, the findings of the present study should be considered speculative, requiring confirmation in wider AHF cohorts. The specific added value of OPP over MAP could not assessed due to collinearity between the two variables. Abdominal organ damage was not assessed, as neither ammonia levels nor more specific parameters of intra-abdominal pressure were collected.²⁷ Echocardiographic non-invasive assessment of CVP was not available retrospectively; however, this study was a proof-of-concept analysis assessing OPP in the context of AHF for the first time and, as such, the ‘gold standard’ invasive evaluation of CVP was deemed suitable and, perhaps, even more appropriate for the purpose of the present work. Should the results of the present study be replicated with the non-invasive echocardiographic assessment of CVP, OPP may truly become a simple bedside tool for risk stratification of AHF patients. MAP and CVP were not recorded continuously; however, the purpose of this analysis was to assess a haemodynamic variable available immediately on admission rather than after a few hours’ recording; also, as continuous medications tailoring is likely to be performed in the Intensive Care Unit rapidly altering the hemodynamic parameters hereby assessed, fluctuations, and mean variations of MAP and CVP detected throughout a long time period may be profoundly confounded by the human intervention (by means of drugs administration and variations) and less reflect the intrinsic individual condition. Continuous pressure monitoring, however, might provide additional information, such as a pressure-time burden or the time that OPP was below a certain threshold. Left ventricular size and ejection fraction were derived from a transthoracic echocardiographic bedside assessment, but data from 3D-echocardiography and cardiac magnetic resonance, being the gold standard techniques for such measurements, were missing. Also, fluid balance data were not collected.

Conclusion

In this retrospective observational study including patients hospitalized for AHF and treated with sodium nitroprusside, OPP on admission was significantly associated with WHF at 48 h at multivariable analysis. Relying on its high sensitivity, OPP could be used as a ‘gatekeeper’ to identify patients at risk of in-hospital deterioration. Future prospective studies evaluating OPP on wider AHF populations are warranted.

Supplementary material

Supplementary material is available at European Heart Journal: Acute Cardiovascular Care.

Acknowledgements

None.

Funding

None.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References

Author notes