NT-proBNP in Stable COPD and Future Exacerbation Risk: Analysis of the SPIROMICS Cohort

Abstract

Background

High N-terminal pro-brain natriuretic peptide (NT-proBNP) during COPD exacerbations is associated with worse clinical outcomes. The prognostic value of NT-proBNP measured during clinical stability has not been well characterized.

Methods

We studied SPIROMICS participants 40–80 years of age with COPD GOLD spirometric stages 1–4. The association between baseline NT-proBNP and incident COPD exacerbations within one year of follow-up was tested using zero-inflated Poisson regression models adjusted for age, gender, race, body mass index, current smoking status, smoking history, FEV₁ percent predicted, COPD Assessment Test score, exacerbation history, total lung capacity on chest CT and cardiovascular disease (any of coronary artery disease, myocardial infarction or congestive heart failure).

Results

Among 1,051 participants (mean age 66.1 years, 41.4% women), mean NT-proBNP was 608.9 pg/ml. Subjects in GOLD stage D had the highest mean NT-proBNP. After one year of follow-up, 268 participants experienced one or more COPD exacerbations. One standard deviation increase in baseline NT-proBNP was associated with a 13% increase in the risk of incident exacerbations (incident risk ratio 1.13; 95% CI 1.06–1.19; p<0.0001). This association was maintained in participants with and without cardiovascular disease.

Conclusion

Baseline NT-proBNP in COPD is an independent predictor of respiratory exacerbations, even in individuals without overt cardiac disease. The impact of detection and treatment of early cardiovascular dysfunction on COPD exacerbation frequency warrants further investigation.

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) exacerbations are associated with poor quality of life (1), accelerated lung function decline (2), increased mortality (3) and elevated healthcare costs (4). Therefore, identification of COPD patients at increased risk of exacerbations has the potential to improve outcomes. Although a history of prior exacerbations is currently the best predictor of future events (5), the variability of exacerbation frequency both within and between COPD subjects (6) and the overall stochastic nature of these events suggest that multiple mechanisms are likely involved. Blood biomarkers, which can easily be measured during clinical stability, are attractive candidates for inclusion in exacerbation prediction models, but their role in this context has yet to be defined.

N-terminal pro-brain natriuretic peptide (NT-proBNP) and BNP are fragments of the precursor hormone pro-BNP which is synthesized in and released from ventricular myocytes in response to myocardial stretch (7). NT-proBNP is a well-established diagnostic and prognostic biomarker in congestive heart failure (CHF) (8, 9) and has also been found to be an independent predictor of death and cardiovascular events in non-ST elevation acute coronary syndrome (10), stable coronary artery disease (11), and even in asymptomatic individuals sampled from the community (12). Natriuretic peptide levels increase during acute exacerbations of COPD and return to baseline after successful treatment (13, 14). Higher NT-proBNP levels measured at the time of a COPD exacerbation have been associated with increased need for intensive care (15) and higher mortality (16–19). However, the role of NT-proBNP as a prognostic biomarker in stable COPD has not been fully investigated. NT-proBNP is a particularly interesting potential marker of risk in COPD because even in patients not experiencing an exacerbation, its levels are higher compared to healthy controls and individuals with asthma (13, 20, 21).

We hypothesized that higher NT-proBNP levels measured during clinical stability are associated with an increased risk of future COPD exacerbations. To test this association, we conducted a prospective analysis of the Subpopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS) cohort.

METHODS

Participants and Study Design

SPIROMICS is a multicenter longitudinal study funded by the National Health Lung and Blood Institute (ClinicalTrials.gov identifier: NCT01969344) designed to identify different COPD subgroups and to validate intermediate outcome measures for use in therapeutic clinical trials (22). It enrolled participants 40–80 years of age between 2010 and 2015 who were either never-smokers (< 1 pack-year smoking history) with pre-bronchodilator forced expiratory volume in 1 second to forced vital capacity (FEV₁/FVC) ≥ 0.7 and FVC > lower limit of normal (23) or current/former smokers (at least 20 pack-year smoking history) with and without airflow obstruction defined as post-bronchodilator FEV₁/FVC < 0.7. Post-bronchodilator spirometry was obtained after each of four inhalations of albuterol 90 µg and ipratropium 18 µg. Spirometric tracings were independently reviewed to confirm that they met American Thoracic Society and European Respiratory Society standards (24). Participants were classified in stages of Global Initiative for Chronic Obstructive Lung Disease (GOLD) spirometric severity (25) as well as in GOLD ABCD stages based on both 2011 and 2017 guidelines (26). In addition, extensive data was collected at the initial study visit, including demographics, comorbidities, questionnaires to assess symptoms and quality of life, cigarette smoke exposure, 6-minute walk distance and high-resolution computed tomography (CT) of the chest (22). Total lung capacity (TLC) was computed from chest CT at full inflation (VIDA Diagnostics, Iowa City, IA). The SPIROMICS protocol was approved by the institutional review boards of all participating institutions and all participants gave written informed consent.

NT-proBNP Levels

The characteristics of blood assays used in SPIROMICS have been previously reported (27). All NT-proBNP measurements described in this analysis were obtained from blood samples collected at baseline using a Myriad-RBM assay (Austin, TX).

COPD exacerbations

Self-reported history of COPD exacerbations in the year before study enrollment was taken from participants at the initial visit. Prospective exacerbation data was collected every three months through a structured telephone questionnaire. Exacerbations were defined as respiratory flare-up events that required health care utilization (office visit, Emergency Department visit or hospital admission) and the use of antibiotics and/or systemic corticosteroids. Severe exacerbations were defined as those necessitating a visit to the Emergency Department or hospitalization. Exacerbations were fully managed by the participants’ usual care providers, with the study investigators not providing any guidance on treatment.

Statistical Analysis

We assessed demographic, spirometric, symptom burden and comorbidity differences between participants with higher (≥ 900 pg/ml) and lower (< 900 pg/ml) baseline NT-proBNP levels, a cutoff chosen for its high accuracy in acute CHF (8). Continuous variables were described as means and standard deviations and categorical ones as percentages. We also examined the distribution of NT-proBNP across GOLD spirometry stages and GOLD ABCD stages based on 2011 and 2017 guidelines. We used two-sample t-tests to compare NT-proBNP levels between individual spirometric stages and individual ABCD stages. The association between baseline NT-proBNP and incident total COPD exacerbations within one year of follow-up was tested using multivariable zero-inflated Poisson models adjusted for age, gender, race, body mass index (BMI), current smoking status, smoking history, FEV₁ percent predicted, COPD Assessment Test (CAT) score (28), prior history of COPD exacerbations, TLC measured on chest CT and cardiovascular disease (coronary artery disease [CAD], myocardial infarction [MI] and CHF). Continuous NT-proBNP was standardized and the associations were presented by change in one standard deviation of the distribution (29); an alternative model using NT-proBNP ≥ 900 pg/ml as a dichotomous variable was also tested. Additionally, we used the same covariates as above to examine the relationship between NT-proBNP and severe exacerbations (≥1) using multivariable logistic regression. We also conducted subgroups analyses according to the presence of underlying cardiovascular disease (defined as CAD, MI or CHF) and tested for interactions between NT-proBNP and cardiovascular disease. All analyses were performed in SAS 9.4 (Cary, NC, USA). A p-value < 0.05 was considered statistically significant.

RESULTS

Description of participants

We identified 1,051 ever-smoker participants with airflow obstruction (post-bronchodilator FEV₁/FVC < 0.7) who had an NT-proBNP level drawn at baseline. A description of the baseline characteristics of participants stratified by NT-proBNP level is presented in Table 1. Subjects with NT-proBNP ≥ 900 pg/ml were older, more likely to be women and less likely to be African-American, had a greater smoking history and were more likely to report CAD, MI and CHF as comorbidities. Those with NT-proBNP < 900 pg/ml had lower FEV₁ percent predicted and greater TLC on chest CT. Mean NT-proBNP in the entire cohort was 608.9 ± 894.1 pg/ml.

Table 1.

Baseline characteristics of study participants

	NT-proBNP < 900 pg/ml (n=849)	NT-proBNP ≥ 900 pg/ml (n=202)	p-value
Demographics
Age	65.2 ± 7.7	70.0 ± 6.8	< 0.0001
Female gender (%)	334 (39.3%)	101 (50.0%)	0.006
African-American (%)	123 (14.5%)	11 (5.5%)	0.0005
Smoking exposure
Smoking history (pack-years)	52.4 ± 23.6	57.3 ± 26.8	0.009
Current smoker (%)	281 (33.1%)	59 (29.2%)	0.29
Markers of respiratory health
Post-bronchodilator FEV1 (% predicted)	61.4 ± 23.8	65.5 ± 20.1	0.02
TLC (L)	5.4 ± 1.3	5.0 ± 1.2	< 0.0001
CAT score	14.9 ± 7.8	14.2 ± 7.9	0.28
6MWD (m)	401.1 ± 118.5	384.3 ± 125.4	0.08
≥ 1 COPD exacerbation in past year (%)	246 (29.0%)	53 (26.2%)	0.44
Comorbidities
CAD (%)	79 (9.5%)	44 (22.0%)	< 0.0001
MI (%)	50 (6.0%)	31 (15.5%)	< 0.0001
CHF (%)	16 (1.9%)	13 (6.5%)	0.0004
Any of CAD, MI or CHF (%)	109 (12.8%)	62 (30.7%)	< 0.0001
GERD (%)	258 (30.4%)	68 (33.7%)	0.37
Obesity (%)	271 (31.9%)	58 (28.7%)	0.38

Data are presented as mean ± standard deviation for continuous variables and as number (percentage) for categorical variables. NT-proBNP, N-terminal pro-brain natriuretic peptide, FEV₁, forced expiratory volume in one second; TLC, total lung capacity measured on chest CT; CAT, COPD Assessment Test; 6MWD, 6-minute walking distance; CAD, coronary artery disease; MI, myocardial infarction; CHF, congestive heart failure; GERD, gastroesophageal reflux disease; obesity defined as BMI > 30 kg/m².

NT-proBNP level by GOLD spirometry and ABCD stages

GOLD spirometry stage 4 subjects (n = 92) had the lowest mean NT-proBNP (338.6 ± 383.4 pg/ml), with a statistically significant difference found compared to subjects in GOLD spirometry stage 2 only, but not those in GOLD stages 1 or 3 (Figure 1). NT-proBNP levels by 2011 and 2017 GOLD ABCD stages are shown in Table 2. In the 2017 classification, 24 subjects moved from stage C to A and 176 subjects moved from B to D compared to the 2011 classification. There was no statistically significant difference in NT-proBNP levels between any two individual GOLD stages in the 2011 classification. In the 2017 classification, subjects in stage D had significantly higher NT pro-BNP levels compared to those in stage B (758.4 vs. 574.9 pg/ml, p=0.03)

Figure 1.

Box plots of NT-proBNP by GOLD spirometry stage

Table 2.

NT-proBNP level by GOLD stages ABCD based on 2011 and 2017 guidelines

	GOLD stages	A	B	C	D
2011 guidelines	n	286	384	56	325
	NT-proBNP level	575.4 ± 652.7	664.3 ± 1032.8	634.5 ± 692.7	568.6 ± 931.3
2017 guidelines	n	310	533	32	176
	NT-proBNP level	580.5 ± 652.0	574.9 ± 728.6	629.1 ± 731.0	758.4 ± 1519.9

NT-proBNP levels are presented as mean ± standard deviation.

Association of NT-proBNP with COPD exacerbations

After one year of follow-up, 268 (25.5%) of the 1,051 participants experienced one or more COPD exacerbations. A higher baseline NT-proBNP was associated with an increased incidence of COPD exacerbations in multivariable analysis of the entire cohort (Table 3); for one standard deviation increase in NT-proBNP, there was a 13% increase in the risk of total exacerbations (Incidence Risk Ratio [IRR] 1.13; 95% CI 1.06–1.19; p < 0.0001). We also examined the relationship between an NT-proBNP threshold of ≥ 900 pg/ml and exacerbations in this same model, and again found a significant association between higher NT-proBNP and exacerbation risk (IRR 1.62; 95% CI 1.19–2.21; p = 0.002). Similarly, one standard deviation increment in NT-proBNP was associated with a higher incidence of severe exacerbations (those requiring Emergency Department visit or hospitalization) during one-year follow-up (IRR 1.27; 95% CI 1.06–1.52; p = 0.009). The incidence risk ratio of severe exacerbations when NT-proBNP ≥ 900 pg/ml was 1.65 (95% CI 0.89–3.06; p=0.11). Finally, we performed additional analyses in participants with and without cardiovascular disease and found that the association between NT-proBNP and incident total COPD exacerbations was maintained in both subgroups (Table 3). An interaction between NT-proBNP and cardiovascular disease was not statistically significant (p = 0.57).

Table 3.

Association of baseline NT-proBNP with COPD exacerbations during one-year follow-up in multivariable analyses

	All subjects n = 1,051	Subjects with cardiovascular disease n = 171	Subjects without cardiovascular disease n = 880
IRR	1.13	1.22	1.10
95% CI	1.06 – 1.19	1.01 – 1.47	1.02 – 1.20
p-value	< 0.0001	0.04	0.01

All entries represent incidence risk ratios (IRR) for exacerbations per one standard deviation increment of NT-proBNP within each group. Models are based on zero-inflated Poisson regression adjusted for demographics (age, gender, race), smoking exposure (smoking history, current smoking status), markers of lung health (FEV₁ % predicted, COPD Assessment Test [CAT] score, exacerbation history at enrollment and total lung capacity on chest CT), body mass index and cardiovascular disease (coronary artery disease, myocardial infarction and congestive heart failure – included only for analysis of the entire cohort, not subgroup analyses).

DISCUSSION

In a large cohort of smokers with COPD, we found that a higher NT-proBNP level measured during clinical stability is associated with an increased risk of COPD exacerbations within one year of follow-up, regardless of the presence of underlying cardiovascular disease.

Participants with GOLD stage 4 spirometry in our cohort had the lowest baseline mean NT-proBNP with a significant difference found compared to those with stage 2, but not stages 1 or 3. Results from several smaller prior studies examining the relationship between natriuretic peptide (BNP or NT-proBNP) levels and GOLD spirometric stage are inconsistent, ranging from higher levels in GOLD stages 3 and 4 (30), no significant difference across GOLD stages (13, 31) and a trend for lower BNP levels at higher GOLD stages (14). One possibility is that, on average, hyperinflation with more severe emphysema reduces cardiac chamber size and stretch (32, 33), thereby decreasing natriuretic peptide levels in subjects with advanced COPD. Another possibility is that survival bias is contributing to lower NT-proBNP levels in individuals with severe COPD as those with high levels could be underrepresented due to death. Regardless, there is still a wide range of NT-proBNP levels in all GOLD spirometric stages; our multivariable analysis, which adjusts for FEV₁ percent predicted, suggests an increased risk of exacerbations when elevated baseline NT-proBNP levels are present.

When participants were reclassified from 2011 to 2017 GOLD ABCD groups, some moved from stage C to A and others from stage D to B. These findings are consistent with prior reports comparing the distribution of COPD subjects across GOLD stages between the two classifications (34). Whereas the exacerbation risk is assessed based on the severity of airflow obstruction and/or exacerbation history in the 2011 guidelines, it is assessed based on exacerbation history alone in the 2017 guidelines. Under the 2017 classification, NT-proBNP was overall higher in stages C and D compared to stages A and B; it was also significantly higher in stage D compared to stage B. These results indicate that COPD individuals with frequent or severe exacerbations in the previous year have a high NT-proBNP level which then becomes an independent predictor of future exacerbations, even when exacerbation history is accounted for.

While a higher NT-proBNP level measured at the time of a COPD exacerbation has been found to be associated with poor short- and long-term clinical outcomes (15–19), its prognostic role when measured in stable COPD has not been well characterized previously. A study of 60 subjects with COPD showed that the time to the next exacerbation was significantly shorter in those with high baseline plasma BNP levels (13). Although this study enrolled a smaller number of participants and evaluated a different yet related outcome, its findings support the presence of an association between higher natriuretic peptide levels and subsequent COPD exacerbations. We have found a relationship between baseline NT-proBNP and exacerbations (both total and severe) in the year immediately following biomarker measurement. This extends a prior report from the SPIROMICS cohort by Keene et al. that found an association between baseline NT-proBNP and severe, but not total, exacerbations when a group of smokers both with and without airflow obstruction was examined (35).

The pathophysiologic mechanisms behind the association we found still need to be determined, but it is likely that cardiac dysfunction both in individuals with and without overt cardiovascular disease contributes to or is uncovered by acute exacerbation events (36). Cardiac comorbidities are more prevalent in subjects with COPD even after accounting for shared risk factors like age and smoking (37, 38). Compared to subjects with COPD and no ischemic heart disease, Patel and colleagues found that those with COPD and ischemic heart disease have higher NT-proBNP levels both in the stable state and during COPD exacerbations as well as a longer symptom recovery time following exacerbations (39, 40). Left ventricular diastolic dysfunction is also common in COPD (41) and can be found in as many as 90% of patients with advanced disease and no cardiovascular comorbidities (42). Brain natriuretic peptides can also be elevated in right ventricular strain and cor pulmonale. A significant positive correlation has been found between natriuretic peptide level and systolic pulmonary artery pressure in stable COPD subjects (20), even among those without congestive heart failure or a history of pulmonary embolism (30). In addition, Wells and colleagues found that an enlarged pulmonary artery (defined as the ratio of pulmonary artery to aorta diameters greater than 1 on chest CT) was independently predictive of total and severe COPD exacerbations (43). Therefore, the presence of underlying subclinical cardiac dysfunction (such as early diastolic heart failure or early pulmonary hypertension) in COPD subjects without overt cardiovascular disease could either increase the severity of an acute COPD exacerbation (thereby escalating the event to require the attention of a healthcare provider) or be the cause of the acute respiratory event itself as more than 20% of exacerbations occur in the absence of pulmonary infection or inflammation (44, 45). Elevation of NT-proBNP reflects diverse aspects of cardiopulmonary stress such as systolic dysfunction, diastolic dysfunction, right heart strain and pulmonary hypertension, which limits the utility of this biomarker when applied in isolation. It remains to be determined whether a subset of COPD patients with higher baseline NT-proBNP and no known comorbid cardiovascular disease would benefit from early evaluation (by echocardiography or cardiac stress test) and treatment of cardiac dysfunction to minimize their risk of adverse events, including respiratory exacerbations.

Our study has a number of strengths such as its large and well characterized cohort, the inclusion of participants with a wide range of severity of airflow obstruction and its adjustment for relevant clinical confounders including factors known to affect NT-proBNP levels such as age, gender and BMI (46). We also acknowledge several limitations. First, our participants were all evaluated in academic medical centers and may not have the same characteristics or outcomes as subjects sampled from the general population. Second, natriuretic peptide levels can be elevated in renal impairment and objective measures of kidney function were not available for inclusion in our analysis. Third, since comorbidities were self-reported by participants, they could have been subject to information bias. However, multiple epidemiologic studies have showed that self-report of cardiovascular conditions is consistently reliable when compared against medical records (47, 48). Fourth, a COPD exacerbation was defined as a respiratory event that specifically required treatment with antibiotics and/or corticosteroids as determined by each participant’s usual care provider. Since diagnostic error on the part of the treating clinician is possible, this definition can potentially result in the misdiagnosis of heart failure exacerbations or other mimics but has the advantage of reflecting real-life clinical decisions and practice.

In summary, higher NT-proBNP measured during clinical stability in subjects with COPD is an independent predictor of respiratory exacerbations within one year of follow-up. Investigating the contribution and interplay of cardiac (elevated pulmonary artery pressure, left ventricular filling impairment, transient coronary ischemia) and pulmonary (emphysema severity, static and dynamic hyperinflation) variables in COPD will enhance our understanding of exacerbation events in susceptible individuals. NT-proBNP can help to risk stratify patients and may contribute to personalize care in the outpatient setting. Whether evaluation and treatment of early cardiac dysfunction decreases respiratory exacerbation frequency and severity in subjects with COPD warrants further study.

Highlights.

• ■NT-proBNP level in stable COPD is highest in GOLD stage D
• ■NT-proBNP in stable COPD is an independent predictor of respiratory exacerbations
• ■This association is maintained even in the absence of cardiovascular disease

ABBREVIATIONS LIST

BMI

Body mass index

BNP

Brain natriuretic peptide

CAD

Coronary artery disease

CAT

COPD assessment test

CHF

Congestive heart failure

COPD

Chronic obstructive pulmonary disease

CT

Computed tomography

FEV₁

Forced expiratory volume in one second

FVC

Forced vital capacity

GERD

Gastroesophageal reflux disease

GOLD

Global initiative for chronic Obstructive Lung Disease

MI

Myocardial infarction

NT-proBNP

N-terminal pro-brain natriuretic peptide

TLC

Total lung capacity