European Journal of Cardiovascular Medicine

Acute Respiratory Distress Syndrome (ARDS) is a life-threatening form of respiratory failure frequently encountered in critically ill patients, with an estimated prevalence of ~10% among intensive care unit (ICU) admissions globally [1]. ARDS commonly arises secondary to conditions such as pneumonia, sepsis, aspiration, or trauma and is pathologically characterized by diffuse alveolar damage and the accumulation of protein-rich inflammatory fluid in the alveoli, resulting from injury to alveolar epithelial and endothelial cells [1]. Despite advances in supportive care, ARDS continues to have a high mortality rate ranging from 30% to 40% [1].

The definition of ARDS has evolved from the American-European Consensus Conference (AECC) criteria in 1994 to the more reliable and valid Berlin Definition established in 2011, which better predicts mortality and guides clinical management [2]. Current treatment strategies focus primarily on lung-protective ventilation, with limited pharmacological interventions available. Given the considerable burden of morbidity and long-term sequelae associated with ARDS, there is a growing emphasis on early recognition, disease stratification, and outcome prediction [1].

Traditional scoring systems like APACHE II and SOFA have been utilized to predict prognosis in ARDS. However, biomarkers offer the potential for greater specificity and clinical utility by reflecting underlying pathophysiological changes and guiding therapeutic decisions [3]. Among these, cardiac biomarkers such as N-terminal pro-brain natriuretic peptide (NT-proBNP) are well established in diagnosing and prognosticating heart failure, and have also shown promise in non-cardiac conditions like pneumonia and COPD [3]. Given the hemodynamic alterations in ARDS—including right ventricular strain and non-cardiogenic pulmonary edema—NT-proBNP may serve as a relevant biomarker in this population.

There are five key roles for biomarkers in ARDS: (1) predicting ARDS in high-risk patients; (2) stratifying disease severity; (3) providing insight into pathogenesis; (4) monitoring therapeutic response; and (5) predicting outcomes [1]. However, limited studies, especially from India, have evaluated NT-proBNP in ARDS patients. Studies by Reddy et al. and Pal et al. provided preliminary evidence of NT-proBNP elevation in ARDS and other non-cardiac illnesses in Indian settings [4,5], but no Indian-specific reference range currently exists. NT-proBNP may help differentiate ARDS from cardiogenic pulmonary edema and indicate disease severity, although this requires further validation.

Our hospital processed 384 NT-proBNP samples in 2021 using the VIDAS® NT-proBNP 2 assay (bioMérieux), which provides rapid results (<30 minutes) via an ELFA-based sandwich immunoassay [6]. While international reference values suggest upper limits of 125 pg/mL (<75 years) and 450 pg/mL (≥75 years), a value above 1025 pg/mL has been considered highly elevated based on Reddy’s study [4].

Exploring the prognostic significance of NT-proBNP in ARDS could provide clinicians with an accessible, rapid biomarker to assess disease severity and guide therapeutic interventions. This study aims to bridge the existing knowledge gap, especially in the Indian context, by evaluating the association between NT-proBNP levels and short-term outcomes in ICU patients with ARDS.

Study Design and Setting

This was a prospective observational study conducted in the Medical and Surgical Intensive Care Units (MDICU and SICU) of GG Hospital, a 166-bed tertiary care centre in Thiruvananthapuram, Kerala. The study period was from February 1, 2022, to December 31, 2022.

Study Population

The study included adult patients (≥18 years) admitted to the ICU who fulfilled the Berlin Definition of ARDS and had NT-proBNP levels measured within 24 hours of diagnosis. Patients were excluded if they had left ventricular ejection fraction ≤50%, had undergone cardiopulmonary cerebral resuscitation immediately before enrolment, or had inconclusive radiological findings inconsistent with ARDS.

Sample Size

The required sample size was calculated based on a 28-day mortality rate of 55.7%, as reported by Lai CC et al. Using the formula n = 4pq/D², with p = 55.7%, q = 44.3%, and D = 20% of p (i.e., 11.14), the estimated sample size was 79 patients, with 80% power.

Outcome Measures

The primary outcome was 28-day mortality from the onset of acute hypoxemic respiratory failure.

Data Collection and Clinical Management

Patient data including age, sex, comorbidities, severity of ARDS, APACHE II and SOFA scores, and fluid balance were prospectively collected. All patients were managed using a standardized institutional protocol based on the ARDSNet guidelines, incorporating lung-protective ventilation, sedation, conservative fluid strategy, vasopressor support, and adjunctive therapies such as prone positioning when indicated.

Follow-Up and Monitoring

Patients were monitored throughout their ICU stay and followed up for 28 days post-diagnosis to document mortality status and the onset of new organ dysfunctions. Clinical status, laboratory values, and treatment parameters were recorded daily.

Statistical Analysis

Descriptive statistics were expressed as means (±SD) or medians (IQR) for continuous variables and frequencies (%) for categorical variables. The Wilcoxon rank-sum test was used for continuous variables and Chi-square test for categorical variables to compare survivors vs. non-survivors. A p-value <0.05 was considered statistically significant.

A total of 80 patients diagnosed with acute respiratory distress syndrome (ARDS) and admitted to the ICU were enrolled in the study. The primary aim was to assess the prognostic utility of plasma NT-proBNP levels in predicting 28-day mortality. By analyzing the association between NT-proBNP concentrations and clinical outcomes, the study sought to evaluate the potential role of this cardiac biomarker in early risk stratification and outcome prediction in critically ill ARDS patients.

Part 1

Table 1: Baseline Characteristics of Study Population (N = 80)

The mean age of the study population was 70.5 years. Survivors were slightly younger (69.36 years) compared to non-survivors (71.76 years), but the difference was not statistically significant (p = 0.111). Although there was a higher proportion of males among non-survivors (60.5%) compared to survivors (47.6%), the gender difference did not reach statistical significance (p = 0.270), suggesting no clear association between age or gender and 28-day mortality in this cohort.

Table 2: ICU Type and Cause of ARDS (N = 80)

Two-thirds of patients (67.5%) were admitted to the Medical ICU, likely reflecting the respiratory etiology of most ARDS cases. A pulmonary cause (such as pneumonia or aspiration) was identified in 54 patients (67.5%), while extrapulmonary causes (such as pancreatitis or trauma) were seen in 26 patients (32.5%). This supports the global epidemiological trend of pulmonary insults being the predominant cause of ARDS.

Table 3: Severity Scores and ARDS Severity vs 28-Day Mortality

Mortality was significantly associated with both higher APACHE II and SOFA scores (p < 0.001). Non-survivors had a median APACHE II score of 22 compared to 10.58 among survivors, and a median SOFA score of 11 vs 5.5, respectively. Additionally, disease severity strongly influenced outcomes—all 17 patients with mild ARDS survived (100%), while 91.4% of patients with severe ARDS (32 of 35) died. This highlights the prognostic utility of clinical scoring systems and the Berlin classification of ARDS severity.

Table 4: Association of NT-proBNP Levels with ARDS Severity

NT-proBNP levels showed a strong correlation with ARDS severity. Among patients with severe ARDS, 94.3% (33/35) had elevated NT-proBNP levels. Conversely, none of the mild ARDS cases had elevated levels. The distribution was statistically significant (p < 0.001), supporting NT-proBNP as a potential severity biomarker in ARDS, possibly reflecting cardiopulmonary stress due to hypoxemia and right ventricular strain.

Part II.  Study Outcomes

1. 28-Day Mortality Rate

Out of the 80 patients enrolled in the study, 38 patients (47.5%) died within 28 days, resulting in a 28-day mortality rate of 47.5%. The remaining 42 patients (52.5%) survived beyond 28 days.

Figure 1 : Comparison of Survivors and Non-Survivors (N = 80)

Nearly half of the ARDS patients admitted to ICU succumbed within 28 days. This high mortality rate underscores the critical need for early risk stratification and prognostic markers in ARDS management.

1. NT-proBNP Levels in the Study Population

The median NT-proBNP level in the total population was 3121 pg/mL (IQR: 906.50–4442.25 pg/mL).

• Among survivors, the median level was 995.50 pg/mL (IQR: 654–2729.25 pg/mL),
• Among non-survivors, the median NT-proBNP was significantly higher at 4330.50 pg/mL (IQR: 3140.25–8240.75 pg/mL).
• This difference was statistically significant (p < 0.001).

Table 6: Comparison of NT-proBNP Levels with 28-Day Mortality

NT-proBNP levels were significantly elevated in non-survivors compared to survivors. This indicates that NT-proBNP may serve as a useful prognostic biomarker in ARDS, with elevated levels correlating strongly with mortality.

1. ROC Curve Analysis for NT-proBNP

A Receiver Operating Characteristic (ROC) curve was generated to evaluate the prognostic utility of NT-proBNP in predicting 28-day mortality. The area under the curve (AUC) was 0.868 (95% CI: 0.790–0.947),The optimal cut-off value determined was 2491.50 pg/mL. At this threshold, the sensitivity was 86.5% and specificity was 76.7%,
indicating excellent discriminative ability.

Figure 2: ROC Curve for NT-proBNP in Predicting 28-Day Mortality

AUC = 0.868 (95% CI: 0.790–0.947), p < 0.001; Optimal cut-off: 2491.50 pg/mL; Sensitivity: 86.5%, Specificity: 76.7%.

NT-proBNP demonstrates high diagnostic accuracy for predicting mortality in ICU patients with ARDS. A value above 2491.50 pg/mL can effectively identify patients at high risk of 28-day mortality, supporting its role in early risk stratification.

ARDS is a leading cause of mortality in the ICU subset of patients worldwide. The Berlin criteria predict mortality in patients with ARDS better than the AECC definition. Using the oxygenation criteria, the absolute value of the area under the ROC for predicting risk of death was only 0.577743. This finding indicates that the initial hypoxemic level is not an optimal predictor of outcome in patients with ARDS. We need other significant factors for predicting outcome. In this context, there is a growing need for a biomarker that predicts mortality. The clinical utility of lymphocyte/neutrophil ratio as a potential molecular marker to evaluate atypical immunosuppression or impairment in patients with ARDS was studied by Song M et al [6]. The follow-up studies on this biomarker were limited. NT-proBNP is a biomarker that is considered to have a potential prognosticative role in ARDS. The prompt identification of the "at risk" subset of patients and early targeted medical management would help to reduce the mortality in ARDS.

Our study aimed to identify an association between plasma NT-proBNP levels and mortality in ARDS. We had several significant findings.

The mean age of sample population was 70.5 (67.65–73.35). The mean age of survivors was 69.36 (66.08–72.64) and non-survivors was 71.76 (66.84–76.69) which was not significant. Most patients admitted to our ICU were of the advanced age group. Our study did not show any association of demographic parameters (age, gender) with adverse outcome. This was in contrast with the study by Lai CC et al [7] that showed that advanced age was associated with mortality.

In our study, majority of patients had medical conditions. This had no significant association with mortality. It concurs with other studies by Lai CC et al [7] and Bajwa EK et al [8]. Most of the patients (67.5%) had a pulmonary cause of ARDS. It had no significant association with mortality. Similar results were shown in the study by Lai CC et al [7]. Matthay, M.A et al [9] recognized that ARDS develops most commonly in the setting of pneumonia (bacterial and viral; fungal is less common), non-pulmonary sepsis (with sources that include the peritoneum, urinary tract, soft tissue and skin), aspiration of gastric and/or oral and oesophageal contents (which may be complicated by subsequent infection), and major trauma (such as blunt or penetrating injuries or burns). Our study also had a fairly similar distribution with trauma patients being the least. Since our center is in the proximity of large public hospitals in the district, we have limited number of trauma cases. Most of these cases get admitted to government hospitals due to medico-legal implications. The number of trauma-related ARDS patients (by large from the younger age group) were relatively less in our study.

In our study, APACHE II scores and SOFA scores were strongly associated with mortality. Lai CC et al [7] and Bajwa EK et al [8] also showed similar results. Moreover, the SOFA and APACHE II scores are well-validated worldwide for prognostication. The SOFA score has six variables, each representing an organ system. APACHE II score is calculated from 12 physiological variables on admission, comprising the acute physiology score, the patient's age, and chronic health status. As the variables increase, the predictive value also increases. Present study compared both AUC of ROC of these scores with that of a single biomarker (NT-proBNP).

Our study showed a strong correlation between severity of ARDS and 28-day mortality. Mortality significantly increased with the severity of ARDS in the original study on the Berlin criteria for ARDS [10]. The LUNG-SAFE study [11] reported that hospital mortality was 34.9% for patients with mild ARDS, 40% for those with moderate ARDS, and 46.1% for those with severe ARDS. Using the Berlin Definition, stages of mild, moderate, and severe ARDS were associated with increased mortality (27%; 95% CI: 24%–30%; 32%; 95% CI: 29%–34%; and 45%; 95% CI: 42%–48%, respectively). This was contrary to the study by Lai CC et al [7] where the mortality rates were: mild (3.0%), moderate (49.0%), and severe ARDS (48.0%) respectively. Hernu et al [12] also showed results similar to Lai CC et al [7]—28-day mortality rates were 27.9% for mild, 22.8% for moderate, and 49.3% for severe ARDS. In our study, other factors that could have contributed to mortality (cumulative fluid balance, troponin I, etc.) were analysed subsequently and found to be non-significant. The relationship between severity of ARDS and mortality is still not clear with multiple studies showing different results. Further studies are needed to strongly validate the current Berlin criteria as an accurate predictor of mortality.

It was found in our study that majority of patients had Type 2 diabetes, systemic hypertension, and dyslipidemia. Type 2 diabetes mellitus and systemic hypertension had significant association with mortality. The previous studies did not determine the association of co-morbidities with mortality. Our study more or less reflects the real-world scenario.

In a study by Zambon M [13], critically ill ARDS patients had a mortality of 40–46%. The study by Lai CC et al [7] had a 28-day mortality rate of 55.7%. In our study, we had a 28-day mortality of 47.5% which is in concordance with most of the literature. The role of comorbidities contributing to unfavourable outcomes remains unexplored in literature [14] as well as the present study.

The median NT-proBNP of study population was 3121 with IQR (906.50–4442.25). The median NT-proBNP of survivors was 995.50 with IQR (654–2729.25) and that of non-survivors was 4330.50 with IQR (3140.25–8240.75). The NT-proBNP levels were significantly higher among non-survivors (p<0.001).

The ROC analysis of our study demonstrated that serum NT-proBNP values had good sensitivity (86.5%) and specificity (76.7%) in predicting 28-day mortality. The study by Lai CC et al [7] also found NT-proBNP as a strong predictor of 28-day mortality. In a study by Bajwa EK et al [8], the results of the ROC analysis demonstrated NT-proBNP values had 80% sensitivity and 51% specificity. When patients were grouped into strata according to levels above and below optimum NT-proBNP, the levels above the cut-off point were significantly associated with increased odds of mortality [15]. Sun et al [16] and Karmapaliotis et al [17] also showed the same results but used AECC criteria to diagnose ARDS and BNP to assess outcomes.

This study demonstrates that elevated plasma NT-proBNP levels are significantly associated with increased 28-day mortality in patients with ARDS admitted to the ICU. NT-proBNP showed strong prognostic utility, with a high area under the ROC curve and good sensitivity and specificity at the optimal cut-off. These findings suggest that NT-proBNP, a readily available and objective biomarker, can serve as a valuable tool for early risk stratification and mortality prediction in ARDS. Incorporating NT-proBNP measurement into initial ICU assessments may aid clinicians in identifying high-risk patients and guiding timely, targeted interventions to improve outcomes.

Conflict of interest: Nil

1. Matthay, M.A., Zemans, R.L., Zimmerman, G.A. et al. Acute respiratory distress syndrome. Nat Rev Dis Primers 5, 18 (2019). https://doi.org/10.1038/s41572-019-0069-0
2. The ARDS Definition Task Force*. Acute Respiratory Distress Syndrome: The Berlin JAMA.2012;307(23):2526–2533. doi:10.1001/jama.2012.5669
3. Song M, Liu Y, Lu Z, Luo H, Peng H, Chen P. Prognostic factors for ARDS: clinical, physiological and atypical immunodeficiency. BMC Pulmonary Medicine. 2020 Dec;20:1-2.
4. Reddy, , Singh, S., Rawat, A. et al. Pro-brain natriuretic peptide (ProBNP) levels in North Indian children with Kawasaki disease. Rheumatol Int 36, 551–559 (2016).
5. Pal R, Rai A, Dhandapani S, Dutta P. Role of NT-ProBNP to differentiate syndrome of inappropriate antidiuretic hormone secretion and cerebral salt wasting syndrome: The conundrum. Neurol India. 2018;66(5):1389-91.
6. Song M, Liu Y, Lu Z, Luo H, Peng H, Chen P. Prognostic factors for ARDS: clinical, physiological and atypical immunodeficiency. BMC Pulmonary Medicine. 2020 Dec;20:1-2.
7. Lai CC, Sung MI, Liu HH, Chen CM, Chiang SR, Liu WL, Chao CM, Ho CH, Weng SF, Hsing SC, Cheng KC. The ratio of partial pressure arterial oxygen and fraction of inspired oxygen 1 day after acute respiratory distress syndrome onset can predict the outcomes of involving patients. Medicine. 2016 Apr;95(14).
8. Bajwa EK, Januzzi JL, Gong MN, Thompson BT, Christiani DC. Prognostic value of plasma N-terminal probrain natriuretic peptide levels in the acute respiratory distress syndrome. Critical care 2008 Aug;36(8):2322.
9. Matthay, Michael A et “Acute respiratory distress syndrome.” Nature reviews. Disease primers vol. 5,1 18. 14 Mar. 2019, doi:10.1038/s41572- 019-0069-0
10. The ARDS Definition Task Force*. Acute Respiratory Distress Syndrome: The Berlin JAMA.2012;307(23):2526–2533. doi:10.1001/jama.2012.5669
11. Bellani, Giacomo et al. “Epidemiology, Patterns of Care, and Mortality for Patients With Acute Respiratory Distress Syndrome in Intensive Care Units in 50 Countries.” JAMA 315,8 (2016): 788-800.
12. Hernu, R. et al. An attempt to validate the modification of the American- European consensus definition of acute lung injury/acute respiratory distress syndrome by the Berlin definition in a university hospital. Intensive Care Med. 39, 2161–2170 (2013).
13. Zambon M, Vincent JL. Mortality rates for patients with acute lung injury/ARDS have decreased over time. 2008;133(5):1120–1127.
14. Park BH, Park MS, Kim YS, Kim SK, Kang YA, Jung JY, Lim JE, Kim EY, Chang J. Prognostic utility of changes in N-terminal pro-brain natriuretic peptide combined with sequential organ failure assessment scores in patients with acute lung injury/acute respiratory distress syndrome concomitant with septic shock. Shock. 2011 Aug 1;36(2):109-14.
15. Determann RM, Royakkers AA, Schaefers J, de Boer AM, Binnekade JM, van Straalen JP, Schultz MJ. Serum levels of N–terminal proB–type natriuretic peptide in mechanically ventilated critically ill patients–relation to tidal volume size and development of acute respiratory distress syndrome. BMC Pulmonary 2013 Dec;13(1):1-7.
16. Sun YZ, Gao YL, Yu QX, Wang J, Xia YH, Lin HY, Zhou SL. Assessment of acute lung injury/acute respiratory distress syndrome using B-type brain natriuretic peptide. Journal of International Medical Research. 2015 Dec;43(6):802-8.
17. Karmpaliotis D, Kirtane AJ, Ruisi CP, Polonsky T, Malhotra A, Talmor D, Kosmidou I, Jarolim P, de Lemos JA, Sabatine MS, Gibson CM. Diagnostic and prognostic utility of brain natriuretic Peptide in subjects admitted to the ICU with hypoxic respiratory failure due to noncardiogenic and cardiogenic pulmonary edema. Chest. 2007 Apr 1;131(4):964-71.