Long-term outcomes after acute hypercapnic COPD exacerbation

Summary Background Non-invasive ventilation (NIV) is used to treat acute hypercapnic respiratory failure (AHRF) in patients with chronic obstructive pulmonary disease (COPD); however, long-term outcomes following discharge are largely unknown. This study aimed to characterize long-term outcomes and identify associated markers in patients with COPD after surviving the first episode of HRF requiring NIV. Methods This study retrospectively analyzed 122 patients, mean age 62 ± 8 years, 52% female and forced expiratory volume in 1 s (FEV1) predicted 30 ± 13%, admitted with an acute hypercapnic exacerbation of COPD and receiving a first-ever NIV treatment between 2000 and 2012. Results A total of 40% of the patients required hospital readmission due to respiratory reasons within 1 year. Persistent hypercapnia leading to the prescription of domiciliary NIV, older age and lower body mass index (BMI) were risk factors for readmission due to respiratory reasons. Survival rates were 79% and 63% at 1 and 2 years after discharge, respectively. A shorter time to readmission and recurrent hypercapnic failure, lower BMI and acidemia on the first admission, as well as hypercapnia at hospital discharge were correlated with a decreased long-term survival. Conclusion Patients with COPD surviving their first episode of AHRF requiring NIV are at high risk for readmission and death. Severe respiratory acidosis, chronic respiratory failure and a lower BMI imply shorter long-term survival. Electronic supplementary material The online version of this article (10.1007/s00508-018-1364-6) contains supplementary material, which is available to authorized users.


Introduction
Chronic obstructive pulmonary disease (COPD) is the third most common cause of death worldwide, resulting in an enormous burden on patients and on healthcare systems [1]. Mortality is increased in both the short-term and long-term period following an acute exacerbation of the disease [2]. Patients with COPD and acute respiratory failure face a high inhospital mortality of 17-49% [3][4][5]8]. Exacerbations become more frequent and more severe as COPD progresses [6]. The course of COPD involves a rapid decline in health status after the second severe exacerbation and high mortality in the weeks following a severe exacerbation [7]. Long-term mortality rate during the 4 years following hospital admission for acute exacerbation of COPD is high (45%) [9]. Severe exacerbations with acute hypercapnic respiratory failure (AHRF) may warrant admission to the intensive care unit (ICU) for non-invasive ventilation (NIV) [10]. Patients surviving such a respiratory crisis are at high risk for readmission and death [5,11]. Associated markers of outcome following hospitalization for exacerbations of COPD were reviewed extensively by Steer et al. in 2010 [12] and Singanayagam et al. in 2013 [2]. In patients requiring mechanical ventilation, the available literature focused almost exclusively on long-term outcome after COPD exacerbations treated with invasive ventilation in patient cohorts derived from the 1990s. Either no or inconsistent potentially predictive markers for long-term outcomes were identified in these studies [13][14][15][16][17][18][19][20][21][22]. Few studies focused on the outcome following primary NIV, the current therapeutic gold standard. In particular, the long-term outcome in COPD patients surviving a hypercapnic exacerbation treated with NIV on the intensive care unit (ICU) has not been given sufficient attention to date. Therefore, the aim of this study was to characterize the long-term outcome after hypercapnic COPD exacerbations requiring NIV in the ICU and to investigate associated markers.

Setting
The Department for Respiratory and Critical Care at the Otto Wagner Hospital is a pulmonary tertiary care center in Vienna, Austria, providing a broad range of pulmonary care including an intensive care unit (ICU). The ICU receives patients from emergency services, affiliated hospitals and from the wards within the Otto Wagner Hospital. The use of NIV for acute hypercapnic exacerbations of COPD is provided only in the ICU and not on regular wards at this hospital. Therefore, all patients acutely requiring NIV are admitted to the ICU. For those patients receiving a prescription of domiciliary NIV following discharge, training in the use of home NIV is provided on our respiratory monitoring unit (RMU).

Patients
The ICU records where searched for patients admitted with a diagnosis of COPD. Patients admitted to the ICU for NIV were included if the following criteria were met: age >18 years at index admission, underlying diagnosis of COPD as judged by the admitting physician, acute hypercapnic respiratory failure (AHRF; pH < 7.35, paCO2 > 45 mm Hg), primary treatment with non-invasive ventilation (with or without secondary endotracheal intubation) and survival to hospital discharge. All patients received standard medical care for COPD exacerbations based on the Global Initiative for Chronic Obstructive Lung Disease (GOLD) clinical guidelines.
Patients were excluded in the presence of the following: cardiopulmonary resuscitation or surgery immediately prior to the admission to the ICU, respiratory failure due to other causes (e. g. pulmonary embolism, acute myocardial infarction, asthma, bronchiectasis, bilateral pneumonia), prior resection of more than one lung lobe, obstructive sleep apnoea (OSAS), obesity hypoventilation syndrome (OHS), active cancer and congestive heart failure (CHF), i. e. history of hospitalization due to CHF, echocardiographic evidence of reduced left ventricular ejection fraction.

Ethics
The study was conducted according to the principles of the Declaration of Helsinki. The ethics committee of the city of Vienna granted an exemption from requiring ethics approval (reference number EK 09-197_NZ). Given the retrospective, observational design of the trial, the need for informed consent was waived.

Study design
The main study objective was to identify prognostic markers for long-term outcome in COPD patients surviving their first episode of HRF requiring NIV. Prognostic markers for predefined endpoints including survival (time between discharge and death), readmission due to respiratory reasons and recurrence of AHRF necessitating readmission to the ICU with a further episode of ventilatory support were sought. The following data sets were collected from the index admission, at hospital discharge and during follow-up: Endpoints during follow-up • Hospital readmission due to respiratory causes (y/n, date), whereas the term "respiratory reasons" was used as a header for any readmission due to an underlying pathology in the respiratory system (e. g. pneumonia, bronchitis, exacerbation of COPD, pulmonary embolism), also including recurrent acute hypercapnic respiratory failure (AHRF) • Death (y/n, date) • Last date seen alive Definitions • Acute hypercapnic respiratory failure (AHRF) = respiratory acidosis: pH < 7.35, paCO2 > 45 mm Hg • Acidemia: pH < 7.35 • Hypoxemia: paO2/FiO2-ratio < 300 • Hypercapnia = hypercapnic respiratory failure: paCO2 > 45 mm Hg • Persistent hypercapnia: paCO2 > 45 mm Hg at discharge The study was conducted as an investigator-initiated retrospective record review. The index admission was confined to a single center. Follow-up data were derived from a total of 11 centers within the Vienna Hos-pital Association (Wiener Krankenanstaltenverbund, KAV), whereby there was no routine outpatient follow-up scheduled for the patients. All eligible patients treated within the study period were included. Survival time was defined by a patient's vital status. All patients not reported as deceased during the followup were considered to be alive.

Data collection
Data were extracted from electronic health records including the local ICU systems (medis, ICUdoc) as well as the city-wide health database (web.okra) of the Vienna Hospital Association. Data were anonymized and entered into an access database by three independent investigators. All the available data in the ICU health records going back to 2000 were used. Data collection was drawn to a close in 2012 when it was felt that enough patients were included to proceed to statistical analysis. It was chosen not to go back further than 2000 because of a shift in treatment paradigms (patients were more readily intubated in the 1990s).
The city-wide health database was only used for follow-up.

Analysis
Data are presented as median and interquartile ranges or as mean and standard deviation (±SD). After completion of data collection, data were cleaned and outliers (>95% confidence interval) were reconfirmed or corrected. Data were exported to a statistical software package (SPSS for Windows, version 15; SPSS Inc; Chicago, IL, USA) for analysis. Univariable data were described using standard methods. Association between risk factors and endpoints were assessed using bivariable and multivariable logistic regression (for dichotomous outcomes) or Cox regression (for timedependent outcomes). In order to address collinearity, the automatic variable selection algorithm of SPSS was used: all potential associated markers (see methods section) were entered in the model but only those variables kept by both the forward and the backward conditional selection algorithm were included in the final models. Kaplan-Meier curves were used to demonstrate survival and readmission characteristics. Statistical significance was defined as p < 0.05.

Results
There were 939 ICU admissions with a recorded diagnosis of COPD during the study period between February 2000 and June 2012. A COPD was regularly found as a comorbidity and not as the main reason for admission, which lead to the exclusion of the majority of cases (817 out of 939 admissions). Reasons for exclusion and respective frequencies are shown in Table 1. Only the index admission of an individual K Long-term outcomes after acute hypercapnic COPD exacerbation 563 original article  (23) Respiratory failure due to other causes 146 (18) AHRF criteria not met 88 (11) Congestive heart failure 81 (10) Primary intubation 78 (10) OSAS 66 (8) Other/insufficient data 52 (6) Death before hospital discharge 45 (6) Active cancer 45 (6) CPR within 7 days prior to ICU admission 28 (3) AHRF acute hypercapnic respiratory failure, OSAS obstructive sleep apnoea syndrome, CPR cardiopulmonary resuscitation, ICU intensive care unit patient was considered for inclusion and repeat admissions were attributed to follow-up data. A total of 122 individual patients were included. Sex was evenly distributed, mean age was 62±8 years and mean BMI was 25±6. Spirometry was available in about half of the patients and indicated an emphasis on severe and very severe COPD stages with a mild degree of chronic hypercapnia. Of the patients 7 (6%) had a record of previous intubation due to e. g. pneumonia or sepsis, whereas none of these 7 patients had a record of previous intubation due to an exacerbation of COPD at this institution. Of the patients 5 (4.1%) were on domiciliary NIV prior to the admission. Demographics and spirometry results are outlined in Table 2.
564 Long-term outcomes after acute hypercapnic COPD exacerbation K original article

Associated markers of readmission and death
Associated markers of readmission to due respiratory reasons with or without hypercapnic respiratory failure Multivariable Cox regression showed the following risk factors to be independently associated with a lower survival: lower BMI, higher age and lower paO2/FiO2 ratio on admission and a higher paCO2 on hospital discharge (Table 4).

Associated markers of overall survival
Multivariable Cox regression showed the following risk factors to be independently associated with a lower survival: lower BMI, lower pH and lower paO2/FiO2ratio on admission and a higher paCO2 on hospital discharge (Table 5).

Discussion
This article presents the first European study with an exclusive focus on long-term outcomes and associated markers in COPD patients surviving an episode of HRF requiring NIV in the ICU. This patient cohort faces a substantial long-term mortality with high rates of hospital readmission and recurrent hypercapnic respiratory failure. Of the patients 40% were readmitted due to respiratory reasons within 1 year. Survival rates were 79% and 63% at 1 and 2 years after discharge, respectively. Cox regression showed the use of domiciliary NIV, age, cachexia and hypoxemia on admission to be correlated with a higher likelihood of readmission due to respiratory reasons. A shorter time to readmission and recurrent hypercapnic failure, cachexia, acidemia, hypoxemia on admission, as K Long-term outcomes after acute hypercapnic COPD exacerbation 565 original article well as hypercapnia at discharge were correlated with an increased long-term mortality. A range of markers as outlined above were found to be associated with long-term outcomes. It is remarkable to find an association between markers of acute physiology and long-term outcomes as shown in the association of the level of acidosis and hypoxemia on admission and lower long-term survival. It is speculated that a more severe derangement of acute physiology points towards exhausted compensatory mechanisms, which might indirectly influence longterm outcome. A range of candidate markers did not show a correlation with long-term markers in this cohort. Most notably the need for secondary intubation, COPD stage, SAPS3 II and TISS 28 scores, days spent on ventilation and length of hospital stay did not show a correlation with long-term outcomes. In this respect it should be emphasized that the sample exclusively comprised survivors of the acute hospitalization episode whereas patients not surviving to hospital discharge were excluded.
Few studies are available in the literature for direct comparison to this study. In a large Canadian study on long-term outcome following the first hospitalization for a COPD exacerbation, mortality rates were 50% at 3.6 years and 75% at 7.7 years [7]. Unsurprisingly, mortality following COPD exacerbation requiring NIV, as reported in this study, is even higher. There is a bulk of studies on in-hospital and long-term outcomes following invasive ventilation in COPD patients, derived from study periods mostly stemming from the 1990s; however there has been a major shift in care towards primary NIV [10] and the prevalence and prognosis of COPD in critically ill patients has changed substantially [23]. The use of NIV has increased significantly over time among patients hospitalized for acute exacerbations of COPD, whereas the need for intubation and in-hospital mortality has declined [24].
A total of three papers were identified with a research question comparable to this study [11,25,26]. The sample size ranged between 93 and 110 patients in these studies. Chu et al. [11] [25] studied 100 patients in an Australian cohort (mean age 70.6 years, 56% male) also receiving wardbased NIV where 56% needed readmission within 1 year and survival rates at 2 and 5 years were 52% and 26%, respectively. Potentially predictive markers that were identified in these studies are listed in table 6 in the online supplementary material. There is little inter-study reproducibility of potentially predictive markers. The only markers that were reproduced in another study were arterial pH prior to initiation of NIV and age in association with long-term mortality. Long-term survival is consistently poor across the 4 studies, with only 1 in 4 patients surviving 5 years following discharge. Given the high morbidity and mor-566 Long-term outcomes after acute hypercapnic COPD exacerbation K original article tality in COPD patients surviving an episode of HRF, it seems mandatory to discuss the prognosis with the patients and to elicit their wishes concerning readmission to ICU and possible reinstitution of ventilatory support.
The present study is limited by its retrospective design and a relatively small sample size. Furthermore, data on smoking status could not be obtained, LTOT prior to admission, length of stay in hospital in the year prior to admission, the Medical Research Council (MRC) dyspnea score and the Katz Index of Independence in Activities of Daily Living score (some of these factors have been identified as associated markers in other studies, as outlined in the online supplementary material). Spirometry within 3 months of admission was available in only half of the patients. Readmissions outside Vienna or in a private hospital within Vienna might have been missed. In an attempt to define this patient cohort as precisely as possible, relatively rigid exclusion criteria were used. It is acknowledged that this might have introduced some bias due to patient selection. The COPD is regularly found with a range of comorbidities that might themselves lead to respiratory failure. It was decided to attribute the cause of respiratory failure not solely to COPD (and therefore to exclude the patient) when there was recorded evidence of an acute illness that by itself may be reason enough to cause respiratory failure. It is acknowledged that it is difficult to adequately draw the line here in the presence of several comorbidities. As stated, all patients requiring NIV are admitted to the ICU at this hospital. The results may not be transferable to hospitals where NIV is also provided on regular wards.
The data point towards a lower BMI and chronic hypercapnia as targets for clinical intervention. Weight gain improves survival in undernourished patients with COPD [27]; however, methods to achieve a reliable weight gain in COPD patients remain elusive. The addition of long-term NIV to standard treatment improves survival of patients with stable hypercapnic COPD when NIV is targeted to greatly reduce hypercapnia [28]. In seems likely that these findings from stable COPD patients would also be observed in patients following acute hypercapnic respiratory failure. Struik et al. randomized 201 COPD patients with prolonged hypercapnia after ventilatory support for acute respiratory failure to either nocturnal NIV or standard care. An improvement in daytime pCO2 and trend to a better quality of life was found but no effect on mortality or readmission rate was seen after 1 year [29]. In patients who used NIV following an admission for an acute exacerbation of COPD (AECOPD) with AHRF, Galli et al. on the other hand found lower readmission rates and an improved event-free survival after 180 days from an index admission compared to patients who did not use NIV post-discharge [30].
In conclusion, there is substantial morbidity and mortality in this patient population. Promising tar-gets for intervention identified in this study, in line with previously published studies, are a low BMI and chronic hypercapnia. The authors believe that these findings should be validated in a prospective cohort leading to predictive models that can be used to facilitate the shared decision-making process.

Conclusion
Patients with COPD surviving the first episode of HRF requiring NIV are at high risk for readmission and death. Distinct risk factors associated with long-term outcomes were identified; however, there is substantial inconsistency among studies published in this field. Results from this study point towards a low BMI and chronic hypercapnia as possible targets for therapeutic intervention.