Infection is an important complication in patients with ST-elevation myocardial infarction (STEMI) undergoing percutaneous coronary intervention (PCI). It increases the mortality even by 10-fold in 30 days (1–3). Early identification high-risk patients for post-acute myocardial infarction (P-AMI) infections, helps to determine a better treatment strategy, avoided aggressive invasive operations, and potentially improve outcomes.

The N-terminal probrain natriuretic peptide (NT-proBNP) is released predominantly from the ventricular myocardium in response to increased ventricular wall stress (4). It is a well-recognized prognostic marker of multiple disease states, including heart failure (5) and chronic obstructive pulmonary disease (COPD) (6–8), which are potentially associated with a high risk of infection. Previous researches focused on the predictive value of NT-proBNP for clinical outcomes in existing sepsis or septic shock that expanding the clinically awareness and utility of NT-proBNP in these infectious stage (9–11). However, the value of NT-proBNP in predicting infection is still lack of knowledge, especially in patients with AMI. Thus, we aim to assess the relationship between the NT-proBNP and infection, and clinical adverse outcomes in patients with STEMI undergoing PCI.

A total of 659 patients were included (mean age: 64 ± 12 years; female, 23.2%). The baseline characteristics are shown in Table 1. Patients in T3 were older, more likely to be a smoker, had higher Killip grade and longer hospital stay, and tend to have a history of hypertension, stroke, diabetes and COPD. A positive association in heart rate, anemia and creatinine levels were noted with increasing NT-proBNP levels, but a negative association in LVEF, estimated glomerular filtration rate, and transradial access was found.

The incidence of infection (10.5, 17.7, 54.5%) and in hospital all cause death (3.7, 4.1, 16.4%) was higher when NT-proBNP increased from T1 to T3 (Figure 2). T1 and T2 have same rates of MACE (15.5% for each) and both lower than that in T3 (43.6%). Patients with infection were more likely to have invasive procedures such as IABP, and adverse clinical events including death, and MACE. In addition, the baseline characteristics of patients with or without infection were compared (Supplementary Table 1). The summary of subtypes of infection were summarized (Supplementary Table 2).

Univariate analysis showed that lgNT-proBNP (OR = 2.16, 95% CI = 1.85–2.53, P < 0.001) was associated with infection. After adjusting for other significant factors, the multiple logistic analysis verified that lgNT-proBNP was still a significant risk factor of infection (adjusted OR = 1.39, 95% CI = 1.12–1.73, P = 0.003, or adjusted OR = 2.16, 95% CI = 1.73–2.70, P < 0.001) (Table 2).

ROC curve analysis confirmed that NT-proBNP could predict in-hospital infection accurately (AUC = 0.774, 95% CI = 0.73–0.82, P < 0.001) (Figure 3). The optimal statistical cutoff point was 3864 pg/mL by using the Youden index. The predictive value of NT-proBNP were evaluated based on the types of infection. The result showed that NT-proBNP had good predictive value for pulmonary infections (AUC = 0.729, 95% CI = 0.63–0.83, P < 0.001) and urinary infections (AUC = 0.741, 95% CI = 0.58–0.90, P = 0.004) (Table 3).

Multiple logistic analysis showed that lgNT-proBNP was a risk factor for MACE during hospitalization (OR = 1.61, 95% CI = 1.35–1.92, P < 0.001). Kaplan-Meier survival curves showed that patients in higher tertiles had worse outcomes (log-rank test, P < 0.001). At follow-up, multivariate cox analysis found that lgNT-proBNP was a risk factor for MACE (adjusted hazard ratio = 1.92, 95% CI = 1.61–2.29, P < 0.001).

To the best of our knowledge, our study was the first to show that NT-proBNP was a good predictor for P-AMI infection in patients with STEMI undergoing PCI, and provided additional information on the association between NT-proBNP and infection. In addition, the current study also further confirmed that NT-proBNP was a hospital and follow-up prognostic factor for these patients.

Although the definition of infection in our study was similar to that of a previous study (1), the incidence of infection was higher in our patients than previous studies (1, 2, 15). The difference in results may be due to the following reasons. First, the risk of infection was proved to increase along with aging (15). Given the different methods of data presentation, it was difficult for us to compare previous studies, but our data showed that the higher number of elderly patients in our study might have caused the high incidence of infection. Second, apart from aging, other risk factors of infection, including accompanying diseases, might later increase the risk of invasive procedures. Patients in our study might have worse baseline conditions, as 16.5% of our patients had Killip grades of III-IV, which was higher than those in other studies (1, 2). Moreover, previous studies have reported that the risk of infection increases with the increasing duration of intensive care unit (ICU) stay (16, 17). Another study showed that the median ICU stay was 1 day in stable patients with STEMI treated in the ICU after primary PCI. Thus, the higher incidence of infection in our study may be due to the prolonged median ICU stay, compared to that of a previous study (4 days vs. 1 day) (18).

Whether NT-proBNP or BNP could predict an infection remained unclear. Two previous studies including patients who underwent lung cancer surgery demonstrated that preoperative BNP levels could predict postoperative complications (19, 20), including pneumonia, and a trend of BNP for predicting postoperative pneumonia was observed. The following other two researches also demonstrated that NT-proBNP or BNP could predict later infections. One study by Bobik et al. (21) showed that preoperative NT-proBNP levels could predict postoperative infection in 31 patients with atrioventricular septal defect. The other study showed that BNP could predict urosepsis after ureteroscopy in patients with unilateral ureteral obstruction (22). However, Attaran and colleagues who investigated 141 patients undergoing cardiac surgery showed that the preoperative BNP level was not significantly higher in patients with existing postoperative infection (23). Despite the differences in the design and the great heterogeneity among these studies, most of them were limited by the sample size; thus, a definite conclusion on whether NT-proBNP could predict infection during hospitalization cannot be made in their studies. Besides, an observational study found that NT-proBNP were able to predict (AUC = 0.72) the concurrent infection at the time of admission in patients with AMI, but in our study, such cohort were excluded (3). This study combined with our findings suggested that NT-proBNP might well predict infection in patients with AMI.

It is well known that NT-proBNP measurements are commonly performed in patients with AMI, and our results strengthen the role of NT-proBNP for these patients and call for an earlier test. The current study highlighted that patients with elevated NT-proBNP levels are at a high risk of developing infection. However, the pathophysiologic mechanisms behind their association remained unclear. First, NT-proBNP was proved to be associated with COPD severity and could predict its exacerbation in both patients with and without cardiovascular diseases (6). Strong connections have also been found between COPD and pneumonia, because they share similar risk factors and present overlap in terms of epidemiology (24). Therefore, these connections supported the predictive role of NT-proBNP for pneumonia. However, we should also notice that NT-proBNP remains similar for predicting infection, after adjusting for COPD. There are other unknown mechanisms beside COPD. Second, patients with high NT-proBNP levels were more likely to suffer from cardiovascular events and death (25, 26), which may increase the use of invasive procedures, such as IABP implantation or mechanical ventilation, subsequently increasing the risk of infection (27, 28). Third, the increasing NT-proBNP levels was a marker that reflecting cardiopulmonary stress, including systolic dysfunction, diastolic dysfunction, pulmonary hypertension and right heart strain (29, 30), might cause a large infarction size or serious damage, which would then later increase the risk of developing pneumonia (31).

There were several limitations in our study. First, it was prospective observational study with small sample size, and some of the baseline characteristics was difference between the NT-proBNP groups. Although multivariate analysis was used to reduce the inherent bias, potential confounders due to unevaluated variables were still possible. The inherent limitations of observational study warrant a well-designed, prospective, multi-centers researches to evaluate the relationship between the level of NT-proBNP and clinical outcomes. Second, we did not monitor the dynamic change of NT-proBNP levels; hence, its dynamic effect on P-AMI infection is unknown. Therefore, further exploration is necessary to identify the clinical impact of risk stratification using the dynamic changes of NT-proBNP for infection. Third, given that procalcitonin measurement is expensive and not routinely used in patients with AMI, we were not sure about its predictive value for infection in our study cohort. Fourth, the reverse causality was still possible. Finally, we only included the patients with STEMI; thus, the predictive value of NT-proBNP for infection in other populations should be demonstrated in future investigations.

In conclusion, The NT-proBNP level is an independent predictor for infection in STEMI patients undergoing PCI. Patients with a high level of NT-proBNP should be considered to undergo effective prophylactic strategies for infection.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the research ethics committee of Guangdong Provincial People's Hospital. The patients/participants provided their written informed consent to participate in this study.

NT and YL: supervisors of the study and guarantee the study data and accuracy. NT and YL: study concept and design. YD and XW: drafting of the manuscript. CD and YL: statistical analysis. All authors acquisition, analysis, interpretation of data, critical revision, and final approval of the manuscript. All authors agreed to submit the manuscript for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.