This was a retrospective single-centre analysis of anonymised hospital registry-based dataset of patients admitted for AHF in the cardiology ICU of the Emergency Department of the University Clinical Centre of Serbia, Belgrade, Serbia, between May 2020 and July 2022. The ICU of the Department of Cardiology in the University Clinical Centre of Serbia is a tertiary level facility, which hospitalises patients with acute/critical cardiovascular disorders. AHF was defined by the presence of symptoms and signs of HF, corroborated by radiology evidence of congestion (Kerly B lines, plueral effusion) and elevated natriuretic peptide levels regardless of left ventricular (LV) ejection fraction (LVEF), in line with the European Society of Cardiology (ESC) guidelines on the management of HF (13). Patients with the first episode of AHF (de novo AHF) and those with a previous history of chronic HF (decompensated chronic HF) were included. Only patients with a primary diagnosis of AHF, according to the International Statistical Classification of Diseases 10th revision (ICD-10) code I50.* were included. Exclusion criteria were as follows: (1) acute coronary syndrome defined according to the ESC guidelines (14, 15); (2) other cardiovascular emergencies complicated by AHF (e.g., infective endocarditis, pulmonary embolism, high grade atrioventricular block etc); (3) patients in cardiogenic or septic shock at the time of admission; (4) patients with confirmed SARS-CoV-2 or Influenza virus infection; (5) patients with evidence of lower respiratory tract infection at admission (6) patients intubated at the time of admission or within the first 48 h. The study protocol was approved by the institutional ethics review board and informed consent was exempt on the basis of a retrospective design.

Hospital acquired pneumonia was defined according to the modified Infectious Diseases Society of America and the American Thoracic Society criteria (7, 9), including radiographic evidence of an inflammatory infiltrate that is new or progressive (on chest x-ray and/or computed tomography), along with at least two of the clinical findings suggestive of infection, namely, new onset of fever (>37.5 C), purulent expectoration, leucocytosis, elevated levels of C-reactive protein (CRP), procalcitonin, fibrinogen and decline in oxygenation. Documentation of ICD-10 codes J15.* and J18.* was also required. Only patients who developed pneumonia at least 48 h after admission and were not intubated at the time of admission were considered.

Data on baseline demographic characteristics (age, sex), vital signs and HF status at admission (heart rate, systolic and diastolic blood pressure, and manifestations of congestion) were collected from the hospital registry-based dataset. HF status was evaluated in accordance with the ESC guidelines, including standard transthoracic echocardiographic examination performed during hospitalisation to confirm structural and functional alterations (13). Based on echocardiographic exam (performed in all patients), HF was classified as HF with reduced ejection fraction, HFrEF (LVEF ≤40%), HF with mildly reduced ejection fraction, HFmrEF (LVEF >41%–49%) and HF with preserved ejection fraction, HFpEF; the latter two categories were pooled together (HFmrEF/HFpEF) (13). The diagnosis of HFpEF was based the ESC guidelines criteria (13), as follows: (1) presence of symptoms and signs of HF (all patients were admitted with symptomatic acute HF), LVEF ≥50% and “ evidence of cardiac structural and/or functional abnormalities consistent with the presence of LV diastolic dysfunction/raised LV filling pressures, including raised natriuretic peptides”. Diastolic dysfunction was diagnosed in the presence of at least 2 of the 4 criteria (16): (1) left atrial volume index >34 ml/m² (or >40 ml/m² in patients with atrial fibrillation); (2) tricuspid regurgitation jet velocity >2.8 m/s; (3) tissue doppler imaging septal e’ < 7 or lateral e’ < 10; and 4. E/e’ > 14. In patients with only one echocardiographic criterion (due to difficulties imposed by performing exam in severely decompensated patients), HFpEF was considered present if LVEF was ≥50% and elevated admission levels of natriuretic peptides (i.e., NT-fragment of pro-B-type natriuretic peptide, NT-proBNP ≥300 pg/ml) were documented (13). Cardiovascular comorbidities, including a history of arterial hypertension (previous diagnosis of hypertension, including treatment with antihypertensive drugs), ischaemic heart disease (previous diagnosis of angina pectoris, prior myocardial infarction, and/or coronary revascularisation with percutaneous coronary intervention and/or cardiac bypass surgery), dilated cardiomyopathy (previous diagnosis of non-ischaemic dilated LV and systolic dysfunction—LVEF <45%), valvular heart disease (aortic stenosis and/or moderate/severe mitral regurgitation documented by echocardiography), atrial fibrillation (diagnosis of paroxysmal, persistent or permanent atrial fibrillation), peripheral arterial disease (previously confirmed by vascular ultrasound exam or angiography), stroke/transient ischaemic attack (as per medical documentation) were collected. Non-cardiovascular comorbidities including type 2 diabetes mellitus (T2DM—previous diagnosis of T2DM, treatment with glucose-lowering medications or newly diagnosed T2DM during hospitalisation), chronic obstructive pulmonary disease (COPD—previous diagnosis of COPD including prescribed treatment with inhaled bronchodilators/steroids/combined inhalers and/or aminophylline/theophylline), chronic kidney disease (CKD—persistent decrease in estimated glomerular filtration rate, eGFR, <60 ml/min/1.73 m² by Cockcroft–Gault equation) and anaemia (haemoglobin <130 g/L in men, and <120 g/L in women) were assessed from the registry-based dataset. A history of hospital admissions for any cause within the 6 months prior to current hospitalisation was collected. New York Heart Association (NYHA) functional class was defined at admission. Routine laboratory analysis along with inflammatory mediators (maximum white blood cell count, CRP, procalcitonin and fibrinogen levels) and cardiac biomarkers (admission NT-proBNP, and high-sensitivity troponin T levels, hsTnT) were collected. Assessment of congestion included clinical evaluation (dyspnoea, pulmonary rales, jugular venous congestion, lower extremity oedema), radiographic signs at admission chest x-ray (Kerley B lines, unilateral/bilateral pleural effusion), and biomarkers (elevated NT-proBNP levels). Data on the use of inotropes/vasopressors (dopamine, dobutamine, noradrenalin), non-invasive and invasive ventilatory support after the development of pneumonia, and fundamental HF medications prescribed at discharge (angiotensin converting enzyme inhibitors, ACEI; angiotensin-1 receptor blockers, ARB; angiotensin-1 receptor neprilysin inhibitor—sacubitril/valsartan, ARNI; beta-blockers; mineralocorticoid receptor antagonists, MRA; and sodium-glucose type 2 inhibitors, SGLT2I) in surviving patients were documented.

The primary study outcome was in-hospital all-cause mortality. Secondary outcomes included: (i) length of hospital stay, (ii) requirement for inotropic support and/or non-invasive/invasive ventilatory support and (iii) prescription patterns of fundamental HF drug classes at discharge, depending on the presence of hospital acquired pneumonia, in patients with HFrEF and HFmrEF/HFpEF.

Expecting a hazard ratio (HR) of 2.2 for the association between hospital acquired pneumonia and all-cause mortality based on previously published data (11), we calculated a minimum sample size of 514 patients, with a power (1-β) of 0.8, and a 2-sided probability of type I error (α) of 0.05 (17). Numerical continuous variables are presented as mean and standard deviation or median and interquartile range (IQR), and categorical variables as absolute numbers and percentages. Difference between variables were compared using the parametric Student t-test, or non-parametric Man–Whitney U-test for numerical variables, and Pearson Chi-square test or Fisher exact probability test for categorical variables, as appropriate. Clinical predictors of the development of hospital acquired pneumonia were analysed in a multivariable logistic regression model, in which clinical variables from Table 1, achieving p-value < 0.05 in univariable logistic regression analysis were entered. In cases of a correlation between predictor variables (e.g., pulmonary rales and dyspnoea, de novo AHF and history of chronic HF, anaemia and haemoglobin, CKD, serum creatinine and eGFR), the variable with a stronger association in the univariable analysis was used in the multivariable model. Inflammatory mediators were not entered as they depicted maximum values during hospitalisation (including those observed in patients with pneumonia). Independent predictors were defined as variables with a persistent significant association with the development of pneumonia (p-value < 0.05) in the multivariable analysis. Cumulative survival rate during hospitalisation in patients with and without pneumonia was assessed with the Kaplan-Meier analysis and compared using the log-rank test. The association between hospital acquired pneumonia and all-cause mortality was analysed in a Cox proportional hazard model adjusted for clinically relevant covariates, including age, sex, NT-proBNP, de novo HF, baseline LVEF (continuous) and other variables listed in Table 1, with a p-value < 0.05 for the association with all-cause mortality in univariable analyses. If a significant correlation between explanatory variables was identified (e.g., de novo HF and a history of HF, systolic and diastolic blood pressure, eGFR and CKD, HFrEF vs. HFmrEF/HFpEF, etc), a variable with a stronger association in univariable analysis was used for adjustment. Time-to-event or time-to-the end of hospitalisation was used to calculate hazard ratios (HR) with the accompanying 95% confidence intervals (CI). The likelihood of the requirement for inotropic or ventilatory support was analysed in a multivariable logistic regression model, which included hospital acquired pneumonia as predictor variable and other clinical variables (with a p-value < 0.05 in univariable logistic regression analysis) were used for adjustment. The likelihood or prescribing HF medications was analysed separately in patients with HFrEF and HFmrEF/HFpEF discharged alive, according to the same principle as described above. All analyses were performed using the IBM SPSS software version 29, and 2-sided p-value < 0.05 was considered statistically significant.

The study included 638 patients admitted to the cardiology ICU for AHF (mean age, 71.6 ± 12.7 years, male 61.9%). Hospital acquired pneumonia was documented in 137 patients (21.5%) after a median of 5.5 days from admission (IQR, 3.0–7.8 days). Microbiological confirmation was available in a subset of patients who subsequently required intubation and mechanical ventilation, in whom positive endotracheal aspirate revealed the following pathogens: Acinetobacter spp, Klebsiella pneumonia, Pseudomonas aeruginosa or Staphylococcus aureus. Baseline clinical characteristics of the entire cohort and according to the presence of pneumonia are presented in Table 1.

Pneumonia occurred more frequently in patients with de novo AHF compared with those with decompensated chronic HF (49.6% vs. 41.9%, p = 0.010), Table 1. Patients who developed pneumonia had more severe evidence of congestion on admission in terms of dyspnoea (97.1% vs. 83.0%), pulmonary rales (89.8%. vs. 67.7%) and radiographically documented Kerley B lines (98.5% vs. 83.8%) and pleural effusions (85.4% vs. 68.6%), all p-values < 0.001, Table 1. Patients with hospital acquired pneumonia more frequently had a history of a previous hospitalisation withing the past 6 months (29.5% vs. 25.5%%, p = 0.022). Mitral regurgitation and a history of stroke/transient ischaemic attack were more prevalent in patients with pneumonia compared to those without pneumonia (mitral regurgitation, 48.9% vs. 26.1%; prior stroke, 24.1% vs. 11.5%, respectively both p-values < 0.001), while there was no difference in other cardiovascular comorbidities, Table 1. Mean LVEF was slightly higher in patients with pneumonia compared to the rest of the cohort (p = 0.050). Non-cardiovascular comorbidities, including T2DM (52.5% vs. 35.9%, p-value < 0.001), COPD (22.6% vs. 10.4%, p-value < 0.001), CKD (44.5% vs. 23.9%, p-value < 0.001) and anaemia (60.5% vs. 43.9%, p-value < 0.001) were significantly more frequent in patients with pneumonia. There was no difference in the smoking status. Patients with pneumonia had higher maximum levels of inflammatory biomarkers, and higher admission levels of serum creatinine and NT-proBNP, Table 1. Admission eGFR and haemoglobin levels were lower in patients with pneumonia, and there were no significant differences in other laboratory values, including hsTnT, Table 1.

Variables significantly associated with the development of hospital acquired pneumonia are presented in Table 2. In multivariable analysis, the development of pneumonia was independently predicted by radiographic evidence of pleural effusion at admission, de novo AHF, presence of mitral regurgitation, a history of stroke/transient ischaemic attack, T2DM and CKD, and increased admission levels of NT-proBNP (≥ the median value of 5,293 pg/ml), Table 2.

A total of 106 (16.6%) patients died during hospitalisation. In-hospital all-cause mortality rates were significantly higher among patients with hospital acquired pneumonia (27.0%) compared with patients without pneumonia (13.8%), p-value < 0.001. Cumulative Kaplan-Meier time-to-event curves in patients with and without pneumonia are presented in Figure 1. After adjusting for age, sex and other clinically relevant covariates, the development of pneumonia was independently associated with a significantly higher risk of in-hospital all-cause mortality (adjusted HR, 2.10; 95% CI, 1.71–2.84; p-value < 0.001).

The median length of hospitalisation was significantly longer in patients with hospital-acquired pneumonia (median, 14.5 days, IQR 9.5–22 days) compared to patients without pneumonia (median, 10 days, IQR 6–16 days), p < 0.001, as presented in Figure 2.

During hospitalisation, inotropes/vasopressors and ventilatory support were required in 37.7% and 26.0% of the total study population; patients with hospital acquired pneumonia were significantly more likely to require either type of the support (inotropes/vasopressors, adjusted odds ratio, OR, 2.31, 95% CI, 2.16–2.81, p-value <0.001; ventilatory support, adjusted OR, 2.11, 95% CI, 1.76–2.79, p-value < 0.001), Table 3.

Prescription patterns of key evidence-based HF medications in patients discharged alive with HFrEF and HFmrEF/HFpEF are presented in Tables 4, 5, respectively. A total of 77.2% and 58.3% of HFrEF patients were prescribed at discharge with renin angiotensin system inhibitors (ACEI/ARB/ARNI) and beta-blockers, respectively, Table 4. Patients who had hospital acquired pneumonia were less likely to receive either of those drug classes compared to patients without pneumonia (ACEI/ARB/ARNI, 43.4% vs. 84.2%, adjusted OR 0.75, 95% CI, 0.51–0.96, p = 0.018; beta blockers, 41.5% vs. 61.7%, adjusted OR 0.56, 95% CI, 0.34–0.90, p = 0.008, respectively), Table 4. MRA and SGLT2I were prescribed in 50.1% and 44.3% of HFrEF patients, and there was no significant difference in the prescription rates between patients with and without pneumonia, Table 4.

Among patients with HFmrEF/HFpEF, ACEI/ARB/ARNI, beta-blockers, MRA and SGLT2I were prescribed in 80.9%, 45.1%, 43.2%, and 26.0% of patients, respectively, Table 5. Patients with hospital acquired pneumonia had a lower likelihood of being prescribed beta-blockers and MRA (beta-blockers, 29.8% vs. 49.4%, adjusted OR, 0.51, 95% CI, 0.38–0.73, p < 0.001; MRA, 34.0% vs. 45.8%, adjusted OR 0.75, 95% CI, 0.52–0.84; p = 0.010, respectively), whilst there was no difference in the prescription of ACEI/ARB/ARNI and SGLT2I, Table 5.

Several limitations of the present study need to be acknowledged. This was a retrospective analysis of a single-centre hospital registry-based data with a limited sample size, which imposes a limitation to the generalisability of our findings. Due to the retrospective design, there is a possibility that some cases of infection other than pneumonia may have been misdiagnosed, although the diagnosis of pneumonia was based on major clinical guidelines and confirmed by documentation of specific ICD-10 codes. We did not document the reasons and circumstances leading to HF decompensation and several clinical characteristics were not systematically recorded and could not be used as covariates in the analyses, including several risk factors for pneumonia (e.g., recent hospitalisation prior to current admission, residence in the nursing home, immunosuppressive disease or therapy), pathogens isolated from microbiological samples, antibiotic regimens used, mechanical ventilation modes and duration, and reasons for not prescribing certain HF medications. Pneumonia was microbiologically confirmed in a subset of patients requiring intubation, with a positive endotracheal aspirate revealing typical gram-negative bacteria and Staphylococcus aureus, which does not rule out other causative microorganisms. Also, data on a history of previous pneumonia, or potential antibiotic use prior to hospital admission were not available, albeit this information could have improved our understanding of the risk of developing pneumonia and antimicrobial susceptibility of the causative pathogens. Furthermore, the present study has not assessed the factors contributing to the development of pneumonia or cardiovascular deterioration during pneumonia. On the basis of these limitations, observations made in the present study should be interpreted as hypothesis generating and may stimulate further prospective evaluation.