The only published national multicenter registry on acute heart failure (HF) in Brazil reported an in-hospital mortality rate of approximately 12% among patients admitted for this condition.¹ This figure stands in stark contrast to the substantially lower mortality rates reported in high-income countries — about 4%.² Poorer outcomes in Brazil may be partly attributed to the preferential admission of more critically ill patients due to limited bed availability.³ Most acute HF-related hospitalizations in Brazil occur in general medical wards of public hospitals, where patient care is typically provided by health care teams without specialized training in cardiology. Accurate risk stratification is particularly important for resource allocation but remains challenging in routine practice. Clinical risk scores developed in research settings are likely to be underutilized by frontline health care teams.⁴ Additionally, well-established prognostic biomarkers for acute HF, such as natriuretic peptides, have yet to be widely adopted in the Brazilian public health care system.⁵

The reality of the hospital affiliated with our academic institution reflects this scenario. It serves as the primary public referral center for both medical and surgical admissions in a region of seven cities in the interior of the state of São Paulo, with an estimated population of 900,000. As part of standard care, natriuretic peptide testing is typically requested at least once during the hospitalization of patients diagnosed with HF. However, access to this test remains inconsistent due to irregular availability within our healthcare setting. This context prompted us to investigate the potential role of C-reactive protein (CRP) as an alternative biomarker for risk stratification in this patient population. There is indeed a sound theoretical rationale for this approach: (1) CRP testing is universally available across healthcare facilities; (2) there is biological plausibility supporting an association between increased CRP levels and adverse outcomes;⁶ and (3) in the outpatient HF population, increased CRP has already been shown to correlate with increased morbidity and mortality.⁷

Compared with patients admitted for acute HF who are directly transferred from the emergency department to the intensive care unit (ICU), those initially hospitalized in general medical wards present a notably more heterogeneous prognosis. We believe that this latter group in particular may benefit from risk stratification based on readily accessible resources. Therefore, this study aims to investigate the prognostic role of increased CRP in patients with a primary diagnosis of acute HF admitted to a general medical ward within the public health care system.

For this analysis, a total of 59 patients were initially assessed. Of these, 11 were excluded because they were first admitted to the ICU before being transferred to the medical ward. An additional 4 patients were excluded due to the absence of CRP measurements within 72 hours of admission to the emergency department, resulting in a CRP measurement rate of 92%. Therefore, the final study sample included 44 patients, with a mean age of 68 ± 16 years, and the majority were male (57%). The average length of stay in the emergency department prior to transfer to the ward was 1.0 (± 0.7) days.

TTE was performed on the 4th (± 2) day of hospitalization and showed an LVEF <50% in 19 of the 44 patients (43%), with a mean LVEF of 34% (± 6). The underlying etiologies identified in this group were ischemic heart disease (30%) and hypertensive heart disease (20%), followed by alcohol and illicit drug use (10%) as well as valvular disease and Chagas disease (5% each).

Among the subgroup with LVEF >50%, 19 patients (43% of total sample) were classified as having heart failure with preserved ejection fraction (HFpEF). In the remaining six patients with LVEF >50%, specific etiologies for HF were identified: three had severe aortic stenosis, one had severe mitral regurgitation, one had a history of corrective surgery for transposition of the great arteries, and one had hypertrophic cardiomyopathy.

Figure 1 shows the distribution of CRP levels in the overall sample, revealing a right-skewed distribution. The median CRP (Q1–Q3) was 2.5 (1.3–6.0) mg/dL. Based on this median value, the sample was divided into two groups: high CRP (>2.5 mg/dL) and low CRP (≤2.5 mg/dL).

Table 1 shows that there were no differences in age or sex between the groups. However, HF classification differed significantly, with a higher proportion of patients with LVEF <50% (and a lower proportion with HFpEF) in the high CRP group compared to the low CRP group.

Another significant difference was observed in smoking history, with a higher prevalence of current or former smokers in the high CRP group compared to the low CRP group. Other comorbidities did not differ between groups, including coronary artery disease, atrial fibrillation (AF), chronic obstructive pulmonary disease (COPD), diabetes mellitus, stroke, and cancer. No participants were undergoing renal replacement therapy at the time of assessment, and creatinine levels measured concurrently with CRP were not significantly different between the high and low CRP groups.

Regarding HF medications, only the use of spironolactone was significantly more common in the high CRP group compared to the low CRP group, likely reflecting the higher proportion of patients with LVEF <50% in that group. Notably, approximately half of the patients were receiving beta-blockers and renin-angiotensin system inhibitors, and 75% were on furosemide (Table 1).

The targeted clinical assessment for HF was performed around the 3rd (±2nd) day of hospitalization. Differences between groups in hemodynamic profiles (A, B, C, and L) were not statistically significant. No patients were receiving vasopressor support while in the ward. Vital signs, including heart rate, respiratory rate, systolic blood pressure, and diastolic blood pressure, were similar between the groups (Table 1).

Despite higher CRP levels, the proportion of patients with a diagnosis of bacterial infection (inferred from antibiotic prescription) was not significantly higher in the high CRP group compared to the low CRP group (Table 1). The distribution of infection sites (lung, urinary tract, skin, bloodstream, and unspecified) was also similar between groups, with pneumonia being the most frequently identified infection.

Regarding in-hospital mortality, the rate was 30% in the high CRP group versus 4.8% in the low CRP group (p = 0.046; Central Illustration). A logistic regression model was constructed to assess the risk of in-hospital death, incorporating the following variables: demographics, comorbidities, HF classification based on LVEF and hemodynamic profile, use of antibiotics, HF medications, creatinine levels, and CRP group (high vs. low).

The model was statistically significant (χ² = 16.415; df = 2; p < 0.001; R² = 0.553) and identified two independent predictors of in-hospital mortality: high CRP (OR = 25; 95% CI: 2-335) and AF (OR = 35; 95% CI: 3-449) (Table 2).

To further clarify the significance of increased CRP in our sample, we analyzed available NT-proBNP data from seven patients in the high CRP group and eight patients in the low CRP group. All measurements were performed within the first three days of hospital admission. NT-proBNP levels exceeded the age-adjusted diagnostic thresholds for acute HF in 87% of patients. Notably, NT-proBNP levels were higher in the high CRP group compared to the low CRP group (Table 1).

In this study conducted at a general hospital within Brazil's public health care system, we found that even among patients with acute HF deemed stable enough for admission to a general medical ward, in-hospital mortality remained high — about 18%. However, the main finding of this study is that CRP measurement at the time of initial assessment appears to be a promising tool for assisting in risk stratification for in-hospital mortality in this population.

CRP is classified as an acute-phase reactant; however, it is well established that increased CRP levels accompany both acute and chronic inflammatory processes, whether infectious or non-infectious, as well as conditions associated with increased metabolic stress.⁹ Its levels correlate both temporally and proportionally with the intensity of the inflammatory response. Among apparently healthy individuals, CRP levels are influenced by factors such as age, sex, and ethnicity.¹⁰ Common prevalent conditions like hypertension and diabetes, along with lifestyle factors, also contribute to variability in CRP levels.⁹

Using the units applied in the present study, CRP values below 0.3 mg/dL are considered normal, while levels between 0.3 and 1 mg/dL are typically classified as low-grade inflammation.¹¹ Values above 1 mg/dL are generally associated with classic inflammatory states.⁹ No single cutoff reliably distinguishes infectious from non-infectious causes of inflammation, nor does it differentiate bacterial from viral infections within the infectious category. Nevertheless, infections — particularly bacterial — are the most common causes when CRP exceeds 10 mg/dL,¹² and especially when levels surpass 50 mg/dL.¹³ In our sample (Figure 1), most participants had CRP levels within the range classified as moderate elevation (1–10 mg/dL),⁹ a range in which the underlying causes are more diverse and may include inflammatory activity directly related to the acute decompensation of HF itself.

To date, the prognostic impact of increased CRP in HF has been demonstrated primarily in patients with chronic HF.^7,14-16 In this context, higher CRP levels are associated with clinical indicators of disease severity (e.g., functional class), ventricular remodeling, and reduced quality of life.⁷ In a cross-sectional analysis of patients with chronic HF, CRP was shown to provide prognostic information regarding mortality and major adverse events independently of BNP levels.⁷

By comparison, the prognostic role of CRP has been less extensively studied in acute HF. Moreover, all previous studies in this area have extended follow-up beyond the hospitalization period.^17-23 In this context, CRP measured at hospital admission has consistently been a strong predictor of major future outcomes, such as death or rehospitalization. However, these outcomes were assessed over varying follow-up periods, ranging from 120 days,¹⁷ to 1 year,^18-19 and even up to nearly 5 years.²⁰ To the best of our knowledge, this is the first study to provide evidence that increased CRP serves as a risk marker for in-hospital mortality in patients with acute HF.

Regardless of the individual contribution of the various mechanisms leading to CRP increase in acute HF, this widely available marker of inflammatory activity may be useful for early risk stratification of in-hospital mortality among patients admitted with decompensated HF. Future studies are strongly needed to confirm our findings in larger cohorts and to further investigate whether the prognostic value of CRP is independent of BNP.