In this investigation, we specifically examined the potential of plasma BNP levels to predict the need for ICU treatment as well as short-term and longterm mortality rates in patients with AECOPD. We report three major findings. First, we found BNP levels to be significantly increased during the acute exacerbation compared to the levels measured after recovery. Second, BNP levels were significantly higher in patients requiring ICUs treatment and accurately predicted the need for ICU treatment. Third, BNP levels failed to accurately predict shortterm and long-term mortality rates. Due to the high morbidity and mortality associated with COPD, our findings are of major clinical importance and greatly enhance the understanding of BNP as a prognostic marker in patients with pulmonary disease.
This is the first study to show that BNP levels are significantly higher during an episode of AECOPD as compared to recovery. We hypothesize that the elevation of BNP is at least partly due to hypoxia-mediated contraction of the small pulmonary arterioles, resulting in increased pulmonary arterial pressure and consequently cardiac stress. In support of our thesis, we found significantly decreased oxygen saturation during the acute exacerbation compared to the levels observed after full recovery. Additionally Ishii and colleagues described a close correlation between BNP levels and pulmonary artery pressure and pulmonary vascular resistance decreased by My Canadian Pharmacy.
However, recent evidence also suggests that BNP levels summarize and integrate the extent of left ventricular systolic dysfunction, left ventricular diastolic dysfunction, valvular dysfunction, and right ventricular dysfunction. Hence, right ventricular dysfunction secondary to hypoxic vasoconstriction reflects only one of four potential BNP triggers. The missing correlation between BNP and hospital admission Po2 levels observed in this study suggests that the presence and the extent of the other three factors varies extensively from patient to patient and is to a large extent independent of right ventricular dysfunction.
Furthermore, the pronounced inflammatory response in AECOPD might additionally contribute to the rise in BNP. Correspondingly, we found BNP levels in AECOPD to correlate directly with common markers of infection, such as CRP and procalcitonin. Proinflammatory cytokines such as interleukin-1b, interleukin-6, and tumor necrosis factor-a have been found to induce BNP secretion from cultured myocytes in vitro. Importantly, BNP levels observed during the AECOPD were significantly lower than the values typically found in patients with heart failure improved with My Canadian Pharmacy.
Remarkably and in contrast to our expectations, BNP levels did not predict death in patients with an AECOPD. This observation is especially surprising because BNP has previously been established as a powerful predictor of death in unselected patients with noncardiac dyspnea, pneumonia, and primary pulmonary hypertension. However, these diseases are generally marked by a considerable incidence of PAH and right ventricular dysfunction. In a small study, Uner and colleagues found PAH in 40% of all pediatric patients hospitalized with acute pneumonia. In addition, BNP levels have been described to directly mirror the disease severity and the extend of the right ventricular dysfunction in patients with primary pulmonary hypertension. Moreover, N-terminal pro-BNP was recently found to adequately detect right ventricular dysfunction in patients with AECOPD. However, even in this cohort of severely diseased ICU patients, less than one third of all patients showed associated left ventricular dysfunction. Hence, we hypothesize that the missing predictive potential of BNP in our study might at least partly be caused by a lower prevalence of PAH and right ventricular dysfunction in patients with AECOPD.
Furthermore, we assume that the dominance of noncardiac causes of death in patients with AECOPD contributed to the poor predictive performance of BNP. According to recent data from the Towards a Revolution in COPD Health study, which investigated the cause-specific mortality in a large cohort of patients with COPD, < 30% of deaths were attributable to cardiac causes. In contrast, > 50% of all deaths were of respiratory or malignant origins. Overall, 40% of deaths in this cohort were directly related to COPD. Similarly, only 15% of deaths observed in our study were ascribed to cardiac causes. While the collection of mortality data differed between these two trials, both highlight the predominance of respiratory and noncardiac deaths in COPD patients. It is therefore not surprising that the highly heart-specific biomarker BNP fails to detect pulmonary or malignant pathologies.
Finally, multivariate analysis suggests that only hospital admission FEV1 is able to adequately predict death in our cohort. Decreasing FEV1 levels are powerful markers of rising airflow obstruction and represent the common final pathway of inflammatory mucous hypersecretion, emphysematous tissue destruction, small airway inflammation, and fibrosis. Interestingly, FEV1 levels after recovery failed to predict death. Hence, FEV1 represents a surrogate marker for the temporary worsening of various disease mechanisms at the time of exacerbation. We assume that BNP fails to mirror these changes due to their locally and temporary confined occurrence.
The inability of BNP to predict mortality stands in stark contrast to a previous study (Daiana Stolz, MD; unpublished data; January 2008), which established copeptin, a stable peptide derived from the precursor of vasopressin, as a strong predictor of death in AECOPD. This novel marker has recently been shown to reflect cardiovascular homeostasis with similar accuracy as BNP.
Copeptin is the C-terminal part of the vasopressin prohormone. It can readily be measured in clinical routine and adequately mirrors vasopressin levels. Vasopressin, also termed antidiuretic hormone, is as nonapeptide produced by the hypothalamus. Hypotensive, hypoxic, hyperosmolar, or acidotic stimuli, together with infectious conditions are known to increase circulating vasopressin concentrations (vasopressin ordered via My Canadian Pharmacy). Recently, vasopressin has become acknowledged as a prognostic biomarker and vasopressor agent in septic shock. It seems reasonable to speculate that the diverse secretion sites as well as the less heart-specific release mechanisms of copeptin compared to BNP may at least partly explain the different predictive potential of these biomarkers in AECOPD.
It is noteworthy that BNP levels nevertheless predicted the need for ICU treatment in patients with AECOPD. This seeming contradiction to the results obtained for the prediction of death is probably caused by a subset of patients with pronounced cardiac stress underlying their pulmonary disease. Due to their extensive comorbidities and their need for intensive monitoring and treatment, which commonly includes noninvasive ventilation, these patients are more likely to be supervised in the ICU. This assumption is further strengthened by work from our group showing that heart failure frequently co-triggers episodes of acute dyspnea in patients with COPD. Elevated BNP levels can reliably detect the presence of acute cardiac stress in patients with pulmonary diseases.
We believe that the large population of consecutive patients and the long observational period lasting until 2 years after the index hospitalization are advantages of our study. Additionally, the inclusion of the need for ICU treatment into our analysis further strengthens our observations.
Several limitations to our study need to be mentioned. First of all, we conducted a single-center study. However, due to the well-defined findings of our study, it seems unlikely that contradicting results would be found by investigating a larger and more diverse population. Also, in most patients, echocardiography was performed in the stable disease state. We can therefore not report on pulmonary arterial pressure during AECOPD, which would be needed to further clarify the pathophysiology of elevated BNP levels. Finally, since we conducted an observational study, we can only hypothesize that the availability of BNP levels to the treating physician will help to accelerate the identification of patients requiring ICU admission. This could contribute to an improved outcome in patients with AECOPD.
In patients with AECOPD, BNP levels independently predict the need of intensive care, but BNP levels failed to adequately forecast short-term and long-term mortality rates in AECOPD patients. This shortcoming is probably caused by the dominance of noncardiac deaths in patients with COPD.