Mid-regional proadrenomedullin as a potential prognostic factor of NIV outcome in AECOPD

Abstract

Background

Early prediction of non-invasive ventilation failure by using clinical parameters and scoring systems is a promising strategy for reducing mortality in acute exacerbation of chronic obstructive pulmonary disease patients. Our aim was to assess predictors of non-invasive ventilation failure in acute exacerbation of chronic obstructive pulmonary disease and the prognostic role of mid-regional proadrenomedullin.

Results

Forty-five patients were enrolled, comprising thirty-six males and nine females with mean age 63.4 ± 9.22 years. Mid-regional proadrenomedullin could not predict non-invasive ventilation outcome, while acute physiology and chronic health evaluation score, blood pH, arterial carbon dioxide and platelet count had predictive value.

Conclusion

Prediction of non-invasive ventilation outcome in acute exacerbation of chronic obstructive pulmonary disease patients is multifactorial and mid-regional proadrenomedullin alone could not predict this outcome.

1 Background

Chronic obstructive pulmonary disease (COPD) is the fourth leading cause of death worldwide. Acute exacerbation of COPD (AECOPD), especially when leading to hospitalization, plays a major role in increasing disease severity, mortality, and raising the economic burden of COPD [1]. Non-invasive ventilation (NIV), the gold standard treatment for AECOPD with acute hypercapnic respiratory failure, has been used in clinical practice since the end of the last century. It reduced the need for intubation, costs and length of hospital stay, and increased survival in these patients [2]. Precise identification of patients who will not benefit from NIV would avoid the unnecessary discomfort of NIV and the dangerous delay of intubation [3].

COPD is associated with systemic inflammation, and several circulating biomarkers are increased during exacerbation, reflecting the spillover of local airway inflammation into the circulation [4]. Adrenomedullin (ADM) is a 52-amino acids peptide that possess a variety of biological actions as immune-modulating, metabolic, vasodilatory and bactericidal actions [5]. The mid-regional fragment of this peptide “Proadrenomedullin” (MR-proADM), is more stable than the active molecule itself, allowing its indirect quantification [6].

Plasma ADM is elevated in various diseases as arterial hypertension, myocardial infarction, heart failure and septic shock [7]. In addition, tissue hypoxia was found to be a powerful physiological stimulus for ADM synthesis [6]. Several studies had explored the role of MR-proADM in COPD exacerbation as a diagnostic biomarker and its ability to predict mortality in patients requiring NIV with conflicting results [4,5,6]. The aim of this study was to evaluate predictors of NIV failure and to assess the ability of MR-proADM to predict NIV outcome in AECOPD patients.

2 Methods

2.1 Study design

This prospective observational study was conducted in the period between December 2019 and August 2021. The study has been approved by the research ethics committee of our institute (No: MD-201-2019) and was conducted in accordance with the Helsinki Declaration. Written informed consent was obtained from all patients.

Inclusion criteria: COPD patients (previously diagnosed in our institute according to global initiative for chronic obstructive lung diseases “GOLD” criteria) with acute hypercapnic respiratory failure due to acute exacerbation. The following were excluded: patients with pneumonia, congestive heart failure and sepsis.

2.2 Sample size calculation

The calculated sample size is 45 AECOPD patients with acute hypercapnic respiratory failure. This sample size will be large enough to detect a 79% specificity of the MR-proADM for predicting the occurrence NIV failure in AECOPD patients with acute hypercapnic respiratory failure [4], at 95% level of confidence and 80% power of the study. Calculation was performed using the following sample size equation that has been described in Buderer [8], and is based on an estimated specificity of 79%, an estimated prevalence of NIV failure 24% [4], and a margin of error of 14%.

$$n = \left[ {\frac{{Z_{ \propto /2} }}{E}} \right]^{2} *\frac{{S_{p} \left( {1 - S_{p} } \right)}}{{\left( {1 - P} \right)}}$$

where n = calculated sample size, Z_α/2 = the critical value of the Z distribution corresponding to the 95% level of confidence (= 1.96), p = estimated prevalence of NIV failure (24%), S_p = estimated specificity of MR-proADM for predicting the occurrence NIV failure in AECOPD patients “the primary end point” (0.79%), E = margin of error (as of 14%).

2.3 Data collection

Medical history was taken including comorbidities and the need for long term oxygen therapy (LTOT), vital signs were assessed on admission and after 1, 4 and 6 h of NIV application. Arterial blood gases (ABG) analysis was done before and 1 h after application of NIV. Laboratory investigations results including: complete blood count, C reactive protein (CRP), serum level of lactate dehydrogenase (LDH), lactate, urea, creatinine, albumin, sodium, and potassium were collected. Acute physiology and chronic health evaluation (APACHE) II score along with dyspnea, eosinopenia, consolidation, acidemia, atrial fibrillation (DECAF) score were calculated for every patient. Serum level of MR pro-ADM was measured using enzyme linked immunosorbent assay (human proadrenomedullin Elisa kit, Sino Geneclon Biotech Co., Ltd, China).

2.4 Statistical analysis

The data was statistically analyzed using Minitab program 17.1.0.0 for windows (Minitab Inc., 2013, Pennsylvania, USA). Continuous data were presented as mean and standard deviation (SD), and categorical data as number and percentage (%), the normality of data was examined using Shapiro–Wilk test. Comparison between the mean of two groups was performed by independent t-test, while Chi square test was used to compare between groups of categorical data. Logistic regression models with stepwise elimination were performed to determine the predictors of NIV failure in the studied cohort, as well factors associated with 30-days mortality. All tests were two-sided, p-value ˂ 0.05 was considered significant.

3 Results

The study included 45 patients, 36 males (80%) and 9 females (20%) with mean age of 63.4 ± 9.22 years. Patients were grouped into 2 groups; group A: NIV succeeded to correct the acute hypercapnic respiratory failure (34 patients, 75.6%) and group B: NIV failed to correct the acute hypercapnic respiratory failure (11 patients, 24.4%).

NIV failure was defined by any of the following: deterioration of the patient's condition, deterioration of conscious level, failure to alleviate symptoms, development of new symptoms or complications (e.g., pneumothorax, sputum retention, nasal bridge erosion), failure to improve or deterioration in ABG parameters, intolerance or failure of coordination with the ventilator [9]. NIV failure in our study was either due to failure to improve ABG parameters or failure to alleviate symptoms.

Mean values of serum MR-proADM levels in the study population were shown in Table 1.

Table 1 Serum MR-proADM level in study population

Demographics of study population were summarized in Table 2.

Table 2 Demographics of study population

Laboratory investigations and clinical scores of study population were shown in Table 3.

Table 3 Laboratory investigations and clinical scores of study population

Comorbidities in our patients included diabetes mellitus and systemic hypertension.

NIV duration had a mean of 2.93 ± 1.98 days, and mean length of hospital stay was 8.77 ± 3.79 days with no significant difference between both groups. Nine patients (20%) died within 30 days.

Increasing APACHE-II score, increasing serum sodium, decreasing platelet count, lower pH on admission and after 1 h of NIV application, increasing PCO₂ after 1 h of NIV application and presence of diabetes mellitus were found to be predictors of NIV failure (Table 4).

Table 4 Predictors of NIV failure in AECOPD patients

Increasing age, higher total leukocytic count and CRP and lower urea were found to be predictors of 30-day mortality (Table 5).

Table 5 Predictors of 30-day mortality in AECOPD patients

Serum level of MR-proADM was found to be related to higher BMI, higher platelet count (Table 6).

Table 6 Factors affecting the level of MR-proADM in AECOPD patients

4 Discussion

During COPD exacerbations, MR-proADM levels increase, suggesting that it could be of special interest in that specific situation [4]. Increased circulating MR-proADM levels might contribute to the prevention of bacterial infection in AECOPD patients. It was also hypothesized that it could potentially promote bronchodilatation in AECOPD patients through inhibiting bronchoconstriction induced by histamine and acetylcholine. This suggests that plasma proADM levels during an exacerbation may reflect the ability or inability of patients to cope with the acute physiologic stress of exacerbation rather than mirroring the severity of the underlying lung disease [5]. So, we tried to assess the prognostic value of MR-proADM on NIV outcome and 30-day mortality in AECOPD patients.

Our results showed that mean values of serum MR-proADM levels (Table 1) did not vary significantly with NIV outcomes. As far as we know this was the first study to evaluate this role for MR-proADM.

In our study, MR-proADM levels correlated with 30-day survival (Table 1). The study [7] found that ADM levels did not differ significantly between survivors and non-survivors, while another study [4] reported that MR-proADM was significantly associated with the risk of poor outcome at 30 days in AECOPD patients.

In our study, age did not correlate with NIV outcome, which agrees with some studies [10, 11] leading to the conclusion that age should not be considered as a limiting factor for NIV treatment, while contradict with one study [12] who found that old age was associated with NIV failure which can be explained by poor mask-fitting, claustrophobia, excessive secretions, intolerance, agitation, and patient/ventilator asynchrony, most of which may be associated with poor respiratory muscle power [13].

Obesity causes reduction of lung and chest wall compliance, elevation of airway resistance, leading to increased risk of alveolar collapse, all of which could make obese patient more vulnerable to NIV failure [14]. Also, obese patients have worse quality of life, reduced 6-min walk distance, increased dyspnea and greater odds of severe AECOPD [15].

In our study there was no significant effect of BMI on NIV outcome. This was in line with some studies [12, 16] that found no relation between BMI and NIV outcome. And we disagree with other study [17] that found that BMI had the highest accuracy for predicting likelihood of NIV failure.

Diabetes mellitus worsens COPD outcomes as it increases length of hospital stay and risk of death during exacerbations, increasing modified medical research council (mMRC) dyspnea scores, and reducing six-minute walking distance [18]. In this study, diabetes mellitus was found to be a predictor of NIV failure (Table 4). Some studies [18, 19] reported similar results, while others [12] found that comorbidities had no predictive value on NIV outcome.

An important factor to be considered in predicting NIV outcome is the degree of acidosis. Furthermore, improvement in pH after few hours of the treatment, along with decrease in the respiratory rate is a predictor for positive outcome [10]. However; there is much debate about the correct cutoff value of pH to choose [18].

Our study (Table 4) showed that lower arterial pH at presentation was associated with increased risk of failure of NIV and the need for intubation. This is in concordance with other studies [10, 13, 18, 21, 22] that reported that pH on admission was highly related to NIV outcome. In addition to baseline, pH values after 1 h of NIV application was proved to be a powerful predictor of NIV outcome as well, in our study and also in some other studies [11, 23]. On the contrary, studies [3, 11, 24] did not identify any relation between pH and NIV failure.

Another variable that was found to be associated with NIV outcome in our study was PaCO₂ after 1 h of NIV application. These results disagree with study [25] that recorded improvement in PaCO₂ after 1 h in success and failure groups. Our results were in line with results of some studies [10, 11, 23] that reported lower values for the PaCO2 in the successful group after one hour of NIV treatment.

One of the beneficial effects of NIV is unloading of the respiratory muscles, decreasing the work of breathing and associated pulmonary hyperinflation so that dyspnea improves and the respiratory rate can fall [18]. In this study, patients in group B showed higher mean respiratory rate at presentation and after 1 h of NIV application than patients in group A (Table 3), however that was not statistically significant. These observations were in line with results of some studies [10, 23, 25].

APACHE II score could predict NIV outcomes (Table 4), as it was expected, as this index includes several factors each of them independently can predict the outcome. One study [22] reported that the lower APACHE II score contribute much to the lower NIV failure. Another study [25] found that patients with late failure of NIV had higher APACHE II score. Multiple studies [11, 20, 21] recorded similar observations.

Patients in group B had higher serum urea than patients in group A (Table 3), however that was not statistically significant. One study [26] stated that lower urea immediately prior to commencing NIV was significantly associated with a successful outcome and that the link between elevated urea and poor outcome is that it reflects a depleted intravascular volume and reduced vital organ perfusion.

Serum lactate was higher in group B than in group A but not of statistical significance (Table 3). One study [17] found higher serum lactate in the failure group but with no statistical significance which totally agree with our results. Another study [3] found that serum lactate had no predictive value for NIV failure.

Recently, elevated RDW was correlated with variety of diseases, including cardiovascular diseases, cerebrovascular diseases, pulmonary embolism, malignancy, diabetes mellitus and others. In addition, RDW is considered a powerful and independent risk factor for mortality. It is believed that an elevated RDW reflects a serious imbalance of homeostasis in the red blood cell caused by impaired erythropoiesis and abnormal erythrocyte survival, due to all kinds of metabolic disturbance [1]. Our study failed to demonstrate any difference between the 2 patient groups regarding RDW, this contradicts with the study [27] that stated that the presence of low RDW value was a predictor of NIV failure.

Other predictors of NIV failure included lower platelet count and higher serum sodium (Table 4), while hemoglobin, TLC, serum albumin, serum creatinine, CRP had no predictive value in NIV outcome. These totally agreed with Miller et al. [26] except for serum sodium.

In this study, 30-day mortality (Table 5) was 20% with increasing age, higher TLC, elevated CRP and lower serum urea as predictors of mortality. Advanced age, low pH, high paCO2, receiving home oxygen therapy and creatinine levels were associated with 30-day mortality in AECOPD patients in one study [28]. While another study [11] found that low BMI < 20 was associated with higher mortality rates in AEOCPD patients admitted in the intensive care unit.

We acknowledge that our study has some limitations, as the relatively small number of patients and being a single-center study that could limit the generalizability of results. Additionally predictive value of MR pro-ADM on NIV failure needs to be further evaluated for more verification of results.

5 Conclusion

One of the most important factors before application of NIV is to assess the probability of its failure, which can worsen the outcome of patients and increase the risk of mortality. Our results identified APACHE II score, platelet count, serum sodium, arterial pH initially and after 1 h of NIV, PaCO₂ after 1 h of NIV and presence of diabetes mellitus as predictors of NIV failure. This may help appropriate patient selection for NIV and close monitoring, if the probability of failure is seen to be high.