Elevated Natriuretic Peptides in Patients With Severe or Critical COVID-19: A Meta-Analysis

We conducted a systematic review of literature published up to April 7, 2020. Data sources included MEDLINE (PubMed and PubMed Central) and early prepublished literature (www.medrxiv.org). Search terms were as follows: BNP or NT-proBNP or natriuretic or laboratory; COVID or COVID-19 or 2019-nCoV or SARS-CoV-2 or SARS-CoV2 coronavirus 2019. We included human studies, retrospective single-center or multicenter studies, case series and case reports, and randomized controlled trials. We excluded duplicate publications, animal studies, systematic reviews and meta-analyses, and studies that did not pertain to the current COVID-19 pandemic or did not present laboratory values. Two authors (B.B. and T.A.) independently analyzed the search results and determined eligibility. A third author (T.K.) was consulted to resolve instances of ambiguity.

Further analysis was done to identify any publications that included patients from the same hospital to avoid double-counting individual patients. Several criteria informed the decision of which studies to include. First, studies were required to have included laboratory values from patients with severe or critical illness and from patients with nonsevere illness. Second, studies that included data from the same hospital were carefully examined for overlapping study period dates. If any 2 studies had overlapping dates, the study with the longer time span and the larger patient sample was preferred over the other. For example, if 2 different studies from the same hospital evaluated BNP values and had overlapping dates, the larger and longer study was included and the other was not. Conversely, if 2 studies from the same hospital had overlapping dates but one study evaluated only NT-proBNP and the other study evaluated only BNP, the study populations were deemed different and both were included.

Illness severity was defined in various ways across studies. Some studies defined severity at the time of admission according to American Thoracic Society guidelines for community-acquired pneumonia; others predefined severity on the basis of various parameters, such as arterial oxygen partial pressure, respiratory rate, or ICU admission.¹⁰  Most of the studies classified severe illness as a respiratory rates greater than or equal to 30 breaths per minute, oxygen saturation less than or equal to 93% (as measured by pulse oximeter), or a ratio of partial pressure of arterial oxygen to fractional inspired oxygen less than or equal to 300 mm Hg. Other criteria used in a few studies included ICU admission, imaging showing multilobular lesions or lesion progression. Despite this heterogeneity, efforts were taken to match patients with similar illness severity across studies. Authors of the selected manuscripts were not contacted to clarify discrepancies or to solicit additional data.

The pooled mean difference was calculated using inverse variance weighted random-effects models. All study results were combined using conversion formulas for estimating means and SDs from nonparametric data, as previously reported.¹¹  Statistical analyses were performed using Review Manager version 5.3.5 (The Nordic Cochrane Centre).

In our literature search, we identified 9 eligible observational studies that included a total of 1,575 patients with COVID-19 (Fig. 1); no prospective studies or randomized controlled trials were identified. All of the included studies represented patients who were treated for SARS-CoV-2 infection at hospitals located in China. Of the 1,575 patients, 342 (21.7%) were identified as having severe illness. Individual study characteristics are presented in Table I.^12–20 

In our analysis, BNP level was significantly higher in patients with severe COVID-19 than in patients with nonsevere COVID-19 (mean difference, 69.56 pg/mL; 95% CI, 1.77–137.35 pg/mL; P = .04, I₂ = 83%) (Figure 2). Similarly, NT-proBNP levels were significantly higher in patients with severe COVID-19 than in those with nonsevere COVID-19 (mean difference, 518.65 pg/mL; 95% CI, 152.40-884.90 pg/mL; P = .006, I₂ = 86%) (Figure 3).

Given that this analysis of patients infected with the SARS-CoV-2 virus was completed early in the pandemic, all eligible publications were from China, the outbreak epicenter. We found that elevated BNP and NT-proBNP levels were significantly associated with severe or critical illness in these Chinese patients. Elevated BNP and NT-proBNP levels identify patients with acute dyspneic episodes who are having exacerbation of congestive heart failure.

Natriuretic peptide levels, which are elevated in response to myocardial stress, can also increase in disorders that indirectly influence the heart, such as pulmonary embolism, hypertension, sepsis, renal failure, and inflammatory and infiltrative diseases.²²  There is ongoing debate as to whether elevated BNP levels in patients with severe COVID-19 are caused by heart failure exacerbations or are the result of new-onset cardiomyopathy. Also unclear is whether cardiac dysfunction is because of direct viral infection or to sequelae of the body's response to the virus. Patients with COVID-19 often experience ARDS, which can itself cause right-heart strain and increase natriuretic peptide levels. In patients without heart failure, elevations in natriuretic peptides may be caused by ventricular dysfunction after an acute coronary event, elevated stress levels, or myocarditis from the SARS-CoV-2 virus itself. Cardiac injury and arrhythmia were found to be more prevalent in patients with severe COVID-19 than in patients with less-severe COVID-19. Fulminant myocarditis also has been described in patients with COVID-19.^23,24 

Angiotensin-converting enzyme 2 (ACE2), which is widely expressed, including in the lung, heart, gut, and vascular endothelium, has been identified as the SARS-CoV-2 receptor and plays an important role in target organ dysfunction, as ACE2 binding by the virus attenuates an organ-protective anti-inflammatory cascade.^25–27  Infection of the cardiac myocytes, where ACE2 is amply expressed, could help explain the pathogenesis of myocarditis seen in COVID-19 and may re-affirm coronavirus-associated heart damage found in current and earlier pathological studies.^26,28  The interplay between SARS-CoV-2 and ACE2 has been noted to increase angiotensin II levels, which in turn increases natriuretic peptide levels.²⁹ 

Cardiac injury in general has been noted in severe cases of COVID-19, with significant increases in troponin levels particularly noted in patients in the ICU vs those not in the ICU.⁸  In one single-center retrospective study, troponin levels were significantly higher in those who died from complications of COVID-19 than in survivors.⁷  In another retrospective study of 416 patients, NT-proBNP levels were significantly more elevated in those with cardiac injury than in those without.³⁰  Given the dearth of studies evaluating BNP and NT-proBNP levels in patients with cardiac injury, it is difficult to ascertain the association between elevated natriuretic peptides and cardiac injury markers. Gao et al showed that patients with severe COVID-19 and high NT-proBNP levels tended also to have higher cardiac injury marker levels.³¹  However, these same patients also displayed other elevated markers indicative of multiorgan system failure, confounding any attempt to definitively trace elevated natriuretic peptide levels directly to cardiac injury. There are several proposed mechanisms of cardiac injury due to SARS-CoV-2 in cases of severe COVID-19, including infection-related myocarditis and type 2 coronary events because of demand ischemia from severe ARDS. Future histopathological and imaging studies may help to unravel the etiology of cardiac injury.

It is possible that patients with severe COVID-19 and elevated natriuretic peptides have preexisting heart failure that makes them more susceptible to deterioration under the overwhelming burden of viral illness itself—presumably in response to a high-output state or to the conspicuous cytokine storm apparent in especially ill patients with COVID-19.^32,33  Activation of the inflammatory system with production and release of proinflammatory cytokines can contribute to decreased ventricular contractility.³⁴  A cytokine profile described as a hyperinflammatory condition, with features akin to hemophagocytic lymphohistiocytosis, develops in many patients with elevated cardiac injury biomarker levels. In such a hyperinflammatory state, cardiac dysfunction may well be part of a broader multiorgan systemic dysfunction, rather than a solitary myocardial event.^35,36  Moreover, the same culprit cytokines could conceivably create a thrombogenic and atherogenic milieu that could provoke a coronary event.³⁷  In a patient with existing heart failure and COVID-19, either condition could influence myocardial durability and thus contribute to the elevation of natriuretic peptides.

The studies in our meta-analysis were largely heterogeneous, with great variation in outcomes among the studies. This reasons for this effect are undoubtedly multifactorial. First and foremost, what constituted a severe case of COVID-19 differed from one study to the next. Whereas some studies defined a severe case as one that required ICU admission, others determined severity according to a specific set of subjective clinical parameters. In Feng's¹⁶  study, patients were classified as having severe COVID-19 if they met 1 of 3 predetermined conditions, including a specified increase in respiratory rate, oxygen desaturation lower than a specified percentage, or a partial pressure of arterial oxygen lower than a specified number. Some studies did not specifically address how severity was classified; others considered COVID-19 to be severe only if the patient ultimately died.

The studies were inconsistent as to when BNP and NT-proBNP levels were measured during a patient's hospital admission, nor was there any way to ascertain the timepoint at which BNP or NT-proBNP levels were measured, given the retrospective nature of the included studies. Patients can initially present with severe illness or can progress to severe illness from an initially less-severe status. Patients who present with less-severe illness may not have their BNP or NT-proBNP measured at all during admission. This potentially introduces bias, in that a large number of patients with less-severe infection may not have BNP or NT-proBNP values counted toward the mean, and thus the mean for these patients may have been underestimated or overestimated. In addition, all studies included in this meta-analysis essentially gave a snapshot view of how natriuretic peptide levels compared between patients with severe vs less-severe COVID-19; there was no indication that any of the patients were individually tracked through the course of their illness. In the future, it would be interesting to see the trends in patients' natriuretic peptide values over the course of their illness while hospitalized, as this may help clarify whether these peptides could predict illness outcomes or disease severity.

Another limitation of the study was that all patients were from China. It would be beneficial to see the outcomes of severity and natriuretic peptide measurements in patients from other countries affected by the SARS-CoV-2 virus. Nonetheless, we consider that this study sheds light on the known and emerging cardiovascular effects of the SARS-CoV-2 virus and confirms that the virus has a clear effect on myocardial durability in patients with COVID-19, even those with nonsevere symptoms.

Our meta-analysis results suggest that natriuretic peptides levels might help to identify patients at high risk for severe COVID-19. The etiopathology of elevated natriuretic peptides in severe COVID-19 is unknown. Upcoming studies from all countries affected by the virus will help to further delineate whether the cause is directly or indirectly of cardiac origin and whether preexisting heart failure has an influence on this disparity. A larger meta-analysis—especially one that includes data from more countries that have been affected by the virus and that elucidates the differences in BNP and NT-proBNP levels classified according to consistent severity criteria—would contribute to a more definitive understanding of the interplay between the heart and COVID-19.