B-Type Natriuretic Peptide Concentrations, COVID-19 Severity, and Mortality: A Systematic Review and Meta-Analysis With Meta-Regression

Alterations in cardiac biomarkers have been reported in patients with coronavirus disease 2019 (COVID-19) in relation to disease severity and mortality. We conducted a systematic review and meta-analysis with meta-regression of studies reporting B-type natriuretic peptide (BNP) or N-terminal proBNP (NT-proBNP) plasma concentrations in COVID-19. We searched PubMed, Web of Science, and Scopus, between January 2020 and 2021, for studies reporting BNP/NT-proBNP concentrations, measures of COVID-19 severity, and survival status (PROSPERO registration number: CRD42021239190). Forty-four studies in 18,856 COVID-19 patients were included in the meta-analysis and meta-regression. In pooled results, BNP/NT-proBNP concentrations were significantly higher in patients with high severity or non-survivor status when compared to patients with low severity or survivor status during follow up (SMD = 1.07, 95% CI: 0.89–1.24, and p < 0.001). We observed extreme between-study heterogeneity (I2 = 93.9%, p < 0.001). In sensitivity analysis, the magnitude and the direction of the effect size were not substantially modified after sequentially removing individual studies and re-assessing the pooled estimates, (effect size range, 0.99 – 1.10). No publication bias was observed with the Begg's (p = 0.26) and Egger's (p = 0.40) t-tests. In meta-regression analysis, the SMD was significantly and positively associated with D-dimer (t = 2.22, p = 0.03), myoglobin (t = 2.40, p = 0.04), LDH (t = 2.38, p = 0.02), and procalcitonin (t = 2.56, p = 0.01) concentrations. Therefore, higher BNP/NT-proBNP plasma concentrations were significantly associated with severe disease and mortality in COVID-19 patients.


INTRODUCTION
A significant number of clinical and demographic factors have been studied in patients with coronavirus disease 19  in regard to their association with specific clinical presentations and measures of clinical severity (1,2). The evidence of an excessive activation of inflammatory and immunomodulating pathways in patients with the more severe forms of the disease, typically characterized by the development of respiratory failure with or without multi-organ dysfunction, have prompted the search for specific biomarkers of inflammation and immuno-activation in order to develop better predictive models to assist with management (3). The increasing evidence of significant alterations of different organs and/or systems in patients with COVID-19 has also led to the investigation of the predictive capacity of additional, organ-specific, biomarkers. For example, the presence of myocardial injury, associated with several cardiac manifestations, including myocarditis, acute coronary syndrome, and arrhythmias, has been welldocumented in COVID-19 patients with or without pre-existing cardiovascular history (4). Notably, cardiac abnormalities in this group are independently associated with an increased risk of mortality (5). While the exact mechanisms involved in the onset and progression of COVID-19 related myocardial injury remain to be elucidated, several circulating markers of myocardial damage, particularly creatine kinase (CK), and troponin, are being increasingly studied in terms of their predictive capacity (6). Another cardiac complication, heart failure, has been observed in about a quarter of patients with COVID-19 and has been associated with an increased risk of adverse outcomes (7,8). The active peptide B-type natriuretic peptide (BNP) and the inactive peptide N-terminal proBNP (NT-proBNP) are both derived from the human BNP precursor proBNP in the ventricular myocytes. The increased secretion of BNP and NT-proBNP from the heart, in response to high ventricular filling pressures, is routinely used as a diagnostic and prognostic marker in heart failure and, by some, as a marker of the size or severity of ischaemic insults (9)(10)(11). However, its biological and clinical role in patients with COVID-19 is not well-established. We addressed this issue by conducting a systematic review and meta-analysis with meta-regression of studies reporting plasma BNP or NT-proBNP concentrations in COVID-19 patients with different disease severity, based on clinical guidelines or need for hospitalization, mechanical ventilation, or transfer to the intensive care unit (ICU), and clinical outcomes, particularly survival status during follow up.

Search Strategy, Eligibility Criteria, and Study Selection
We conducted a systematic literature search, using the terms "brain natriuretic peptide" or "BNP" or "NT-proBNP" or "N-terminal pro-brain natriuretic peptide" and "coronavirus disease 19" or "COVID-19, " in PubMed, Web of Science and Scopus, from January 2020 to January 2021, to identify peerreviewed studies reporting BNP/NT-proBNP concentrations in COVID-19 patients according to disease severity and/or mortality. We accessed the references of the retrieved articles to identify additional studies. Eligibility criteria were (a) reporting continuous data on plasma BNP/NT-proBNP concentrations in COVID-19 patients, (b) investigating COVID-19 patients with different disease severity or survival status during follow up, (c) adult patients, (d) English language, (e) >10 patients, and (f) full-text available. Two investigators independently screened individual abstracts. If relevant, they independently reviewed the full articles (PROSPERO registration number: CRD42021239190). We used the Newcastle-Ottawa scale to assess study quality, with a score ≥6 indicating high quality (12).

Statistical Analysis
We calculated standardized mean differences (SMD) and 95% confidence intervals (CIs) in BNP/NT-proBNP concentrations between COVID-19 patients with low vs. high severity or survivor vs. non-survivor status. A p < 0.05 was considered statistically significant. When studies reported medians and interquartile ranges (IQR) the corresponding means and standard deviations were estimated (13). We assessed betweenstudy heterogeneity in SMD values using the Q-statistic (significance level at p < 0.10). Inconsistency across studies was evaluated using the I 2 statistic, where I 2 < 25% indicated no heterogeneity, between 25 and 50% moderate heterogeneity, between 50 and 75% large heterogeneity, and >75% extreme heterogeneity (14,15). Random-effect models were used to calculate the pooled SMD and 95% CIs if significant heterogeneity was present. In sensitivity analyses, the influence of individual studies on the overall effect size was assessed using the leave-one-out method (16). The presence of publication bias was assessed using the Begg's and the Egger's test, at the p < 0.05 level of significance (17,18), and the Duval and Tweedie "trim and fill" procedure (19). To identify factors contributing to the between-study variance, we investigated the effects of several biologically and/or clinically plausible factors on the SMD by univariate meta-regression analysis.

Literature Search and Study Selection
We initially identified 1,815 studies. A total of 1,758 studies were excluded after the first screening because they were duplicates or irrelevant. Following full-text revision of the remaining 57 articles, 13 were further excluded because they did not meet the inclusion criteria. Thus, 44 studies in 18,856 COVID-19 patients, 14,569 (53% males, mean age 48 years) with low severity or survivor status and 4,287 (59% males, mean age 61 years) with high severity or non-survivor status, were included in the final analysis (Figure 1 and  reported BNP or NT-proBNP concentrations within the first 24-48 h from admission.

Meta-Analysis
The overall SMD in BNP/NT-proBNP concentrations between COVID-19 patients with low vs. high severity or survivor vs. non-survivor status is shown in Figure 2. In two studies, patients with high severity or non-survivor status had significantly lower BNP/NT-proBNP concentrations when compared to those with low severity or survivor status (mean difference range, −0.45 to −0.67) (31, 59). By contrast, in the remaining studies BNP/NT-proBNP concentrations were lower in patients with low severity or survivor status (mean difference range, 0.02 -5.09), with a non-significant difference in five studies (35,38,44,50,63). Pooled results confirmed that BNP/NT-proBNP concentrations were significantly higher in patients with severe disease or nonsurvivor status (SMD = 1.07, 95% CI: 0.89 -1.24, and p < 0.001; Figure 2). There was extreme between-study heterogeneity (I 2 = 93.9%, p < 0.001). BNP/NT-proBNP concentrations remained significantly higher (SMD = 1.06, 95% CI: 0.86 -1.26, and p < 0.001; I 2 = 93.0%, p < 0.001) in patients with high severity or non-survivor status after excluding two relatively large studies, accounting for nearly 56% of the overall sample size (26,37). In sensitivity analysis, the magnitude and the direction of the effect size were not substantially modified after sequentially removing each study and re-assessing the pooled estimates (effect size range, 0.99 -1.10; Figure 3). No publication bias was observed with the Begg's (p = 0.26) and Egger's (p = 0.40) t-tests. However, using the trim-and-fill method, we identified three potential missing studies to be added to the left side of the funnel plot to ensure symmetry (Figure 4). The adjusted SMD, albeit attenuated, remained significant (SMD = 0.90, 95% CI: 0.70 -1.09, and p < 0.001).

DISCUSSION
In our study, plasma concentrations of BNP and NT-proBNP, generally measured within the first 24-48 h from admission, were significantly higher in COVID-19 patients with severe disease, based on clinical assessment or the need for hospitalization, mechanical ventilation, or ICU transfer, and in those who did not survive when compared to patients with mild disease or who survived during follow up. The observed SMD values for combined natriuretic peptide concentrations or BNP and NT-proBNP separately, 1.07, 1.22, and 0.98, respectively, suggest a biologically and clinically significant effect size (64). Although between-study heterogeneity was extreme, in sensitivity analysis the effect size was not influenced when individual studies were sequentially removed. The Begg's and Egger's t-tests did not show any evidence of publication bias. In meta-regression analysis, significant associations were observed between the SMD value and D-dimer, myoglobin, LDH, and procalcitonin, but not with age, gender, AST, ALT, creatinine, troponin, CK, albumin, ferritin, CRP, WBC, diabetes, hypertension, or cardiovascular disease.
Differently from the inactive NT-proBNP, the BNP exerts significant biological effects through its binding to the guanylyl cyclase-coupled natriuretic receptors A and B. The consequent increase in cyclic guanosine monophosphate causes vasodilatation, diuresis, natriuresis, inhibition of the renin-angiotensin-aldosterone system, inhibition of fibrosis, hypertrophy, cell apoptosis and inflammation, including suppression of superoxide generation by neutrophils, and improvement in myocardial relaxation (10). Notably, there is no evidence of significant associations between BNP and cyclic guanosine monophosphate concentrations in human studies. Furthermore, specific BNP-mediated protective effects, particularly the suppression of neutrophil-mediated generation of superoxide via nicotinamide adenine dinucleotide phosphate oxidase, are impaired in the context of acute heart failure, even in the presence of increased BNP concentrations (65). Whilst such effects are partially restored with pharmacological treatment, the failure of BNP-related suppression of superoxide release might lead to sustained tissue inflammation in heart failure, with or without concomitant COVID-19. There are other differences between the BNP and the NT-proBNP, with the latter being characterized by a higher molecular mass, a longer half-life (>60 vs. 15-20 min), a higher degree of in vivo glycosylation, and a lower degree of intra-individual biological variation (66). The better analytical characteristics of the available immunoassay methods for the measurement of NT-proBNP concentrations, when compared to those for the assessment of the BNP, have prompted some experts to advocate the measurement and monitoring of NT-proBNP concentrations as the best strategy for the management of patients with heart failure (66). These issues notwithstanding, in our meta-analysis the studies reporting BNP vs. NT-proBNP plasma concentrations had similar SMD values and degrees of heterogeneity.
The significant association observed between plasma BNP/NT-proBNP concentrations, disease severity and mortality in patients with COVID-19 is likely to reflect the presence of heart failure and its adverse sequelae in this group. In this context, these routine and relatively inexpensive biomarkers might assist the clinician with the early diagnosis of cardiac dysfunction and the prompt initiation of appropriate pharmacological and non-pharmacological therapies (67). Further research is warranted to determine whether the assessment of BNP/NT-proBNP might also be incorporated into predictive tools that are specifically developed and validated in COVID-19 patients. The reported associations, in meta-regression analysis, between the SMD of BNP/NT-proBNP and D-dimer, myoglobin, LDH, and pro-calcitonin suggests that the effect size is particularly correlated with markers of pro-coagulant activity, skeletal muscle and other tissue damage, and severe sepsis, respectively. Notably, these markers have, in turn, been shown to have significant associations with COVID-19 severity and outcomes (68)(69)(70). By contrast, the lack of associations observed with other cardiac biomarkers, e.g., troponin and CK, suggests that the measurement of BNP/NT-proBNP may provide complementary, rather than redundant, information regarding the presence of cardiac abnormalities in patients with COVID-19.
The extreme between-study heterogeneity represents a limitation of our study. However, we did not observe significant publication bias and the overall effect size was not substantially affected in sensitivity analyses. The lack of significant associations between the SMD and several patient and study characteristics, except for D-dimer, myoglobin, LDH, and procalcitonin concentrations, suggests that other unreported factors might have contributed to the observed heterogeneity. Such factors may include the relationship between the SMD values and the presence of new onset vs. acute on chronic heart failure and the information regarding the specific analytical methods used for the determination of BNP and NT-proBNP plasma concentrations (66). In this context, the lack of available information on indexes of left ventricular function prevented the conduct of further meta-regression analyses of the association between such indexes and the SMD values. Furthermore, virtually all selected studies reported isolated, rather than serial, measurements of natriuretic peptide shortly after hospital admission. This issue is particularly important as the routine monitoring of BNP/NT-proBNP concentrations has been shown to be beneficial in heart failure (71). Further studies should investigate the prognostic value of single vs. serial BNP/NT-proBNP measurements also in patients with COVID-19. In conclusion, higher plasma concentrations of BNP or NT-proBNP are significantly associated with higher disease severity and increased mortality in COVID-19. Additional studies are required to determine whether these cardiac biomarkers can be incorporated into robust predictive tools that further assist with early management and monitoring in this patient group.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.