UK National COPD Resources and Outcomes Project 2008: Patients with Chronic Obstructive Pulmonary Disease Exacerbations Who Present with Radiological Pneumonia Have Worse Outcome Compared to Those with Non-Pneumonic Chronic Obstructive Pulmonary Disease Exacerbations

Chronic obstructive pulmonary disease (COPD) is a common cause of chronic disability and death, there being approximately 25,000 COPD-related deaths annually in the UK [1]. It is the second most common cause of hospital emergency admissions with approximately 100,000 admissions in 2000 [2], associated with 1 million hospital-bed days in England alone. Previous Royal College of Physicians/British Thoracic Society national audits of COPD have already highlighted that COPD patients usually present to hospital with acute exacerbations: these are serious (14% of these patients die during admission) and nearly a third will be readmitted within 90 days of the initial admission [3,4,5].

Community-acquired pneumonia (CAP) is also a common cause of hospital admission, and one of the leading causes of death in the UK [6]. In a subanalysis of a randomised clinical trial, where a total of 262 patients with CAP included 95 (36.3%) who had COPD, it was recently reported that COPD patients do not differ from non-COPD patients in terms of 30-day mortality outcome in CAP [7]. However, there is a distinct paucity of information with regard to the outcome of pneumonia among COPD patients. Data on how patients with COPD exacerbation presenting with pneumonia are managed compared to those without pneumonia is limited, as is comparative data about prognosis adjusting for case mix and individual prognostic indicators.

In this study, we set out to examine the characteristics of patients with COPD exacerbation who presented with radiological pneumonia compared to other COPD admissions using national audit 2008 data [8]. We were interested in comparing the management and outcome between those with and without pneumonia in COPD exacerbation (as opposed to pneumonia patients with COPD) with regard to mortality (inpatient and 90-day) and length of stay, controlling as far as possible for case mix and individual prognostic factors.

The 2008 audit-collected data in 5 different areas of COPD care and all details can be viewed at the Royal College of Physicians’ website (http://www.rcplondon.ac.uk/clinical-standards/ceeu/Current-work/ncrop/Pages/audit.aspx). All UK and island acute trusts admitting COPD exacerbation patients were invited to participate in the audit programme and to collect data as acute units. The term ‘unit’ was used to describe each participating organisation and was defined as ‘a hospital that admits acute unselected emergency admissions’. Thus, where a whole trust participated in the audit, the term ‘unit’ refers to that trust. Where a hospital participated as part of a trust, the term ‘unit’ refers only to that hospital within the trust. Participants were asked to define ‘units’ in terms of the functionality of their respiratory medicine departments.

Each participating unit completed a retrospective case note audit of up to 60 consecutive patients with admissions identified prospectively between March and May 2008. The audit-collected items related to patient characteristics, clinical features at the time of admission and the process of care and clinical outcomes within 90 days of the index admission. The reliability of the clinical case data was assessed by asking units to double-enter data on the first 5 cases using a different auditor. The levels of reliability for 952 submitted cases were generally good, with kappa values of 0.60 and higher dominating, with many over 0.80 (very good). Data were entered onto a bespoke web-based programme and collated centrally at the Royal College of Physicians, London.

Eligible patients were divided into 2 groups, those with a chest X-ray (CXR) shadow consistent with pneumonia associated with COPD exacerbation and those without. The audit data relevant to this classification came from an audit question relating to the CXR appearance within the first 24 h. The precise instruction to the auditor was as follows: ‘Where the patient has had more than one chest X-ray within the first 24 h please refer to the first X-ray. Record all abnormalities detected in the patient’s chest X-ray as documented by the most senior member of staff commenting on the X-ray. Where comments are made describing a number of potential problems NOT seen e.g. no evidence of pneumothorax, then please enter ‘‘no abnormalities’’. Any comment about chest X-ray not included within current options should be recorded as ‘‘other abnormality’’.’ The 8 coding options open to the auditor were: (1) CXR – no abnormality, (2) changes consistent with COPD, (3) changes consistent with pneumonia, (4) suspected or definite cancer, (5) other abnormalities, (6) X-ray – poor quality and unhelpful, (7) no comment made, (8) not taken.

For the purpose of this study, cases were excluded from analysis if a CXR had not been taken, if cancer was suspected or definitely indicated by the X-ray, or if lung cancer was stated as a co-morbidity. Missing data values affect patient denominators and results presented as percentages will vary in their denominators accordingly. Data were analysed using SPSS version 15.0 and STATA 8.0. The χ² test was used to indicate the statistical association of the patient and the clinical features with pneumonia (tables 1, 2). Binary regression methods (STATA ‘binreg’ software) were used to adjust for hospital clustering effects, to obtain p values for pneumonia in relation to outcomes (table 3) and to obtain risk ratios and 95% confidence intervals for pneumonia (table 4). Ethics approval was given by the University College Hospital/ University College London Multicentre Research Ethics Committee.

Clinical data for 9,716 patients were received from 232 acute units within 177 acute NHS hospital trusts (96% of 184 eligible trusts). The median number of cases contributed by units was 46, inter-quartile range 29–58. A total of 378 patients were excluded either because a CXR had not been taken, cancer was suspected or definitely indicated by the X-ray, or lung cancer was stated as a co-morbid condition. Therefore, a total of 9,338 patients were included in the current report.

Within 24 h of admission, 16% (1,505) had changes consistent with pneumonia indicated on CXR, unit median 14% inter-quartile range 9–23%. The overall mean (SD) patient age was 73 (10) years and half (n = 4,710) were male. The mean (SD) FEV₁ % predicted was 41% (19%) for the 5,045/9,338 for whom it was recorded. Where indicated, 98% (8,343/8,517) of the patients were described as ‘white’. Four percent (413) lived in sheltered accommodation and 5% (482) in a residential home, whilst 90% (8,443) lived alone or with someone else in a flat or house. Thirty-nine percent (3,597) received some form of personal care. In-hospital mortality was 7.5% (702/9,338) and mortality within 90 days of the admission date was 13.4% (1,198/8,945).

Baseline characteristics of COPD patients with and without radiological pneumonia are shown in table 1. As a consequence of the large audit sample, small clinical differences reach statistical significance. The strongest correlates (p < 0.001) with pneumonia were age, disability, albumin and urea levels, and the number of co-morbidities. Patients with pneumonia were older and had more limited levels of activity overall, with lower blood albumin and higher blood urea, with a higher prevalence of multiple co-morbidity (≥3 co-morbidities, 23 vs. 19%).

In terms of clinical management (table 2), pneumonia patients were much more likely to receive antibiotics within 24 h (p < 0.001). There were lesser marginal differences in regard to treating with steroids, oxygen prescription and taking arterial blood gases. Pneumonia patients were more likely (p < 0.001) to have a lower bicarbonate level and PaO₂ of <7.3, but no significant differences were observed between the 2 groups with regard to the prevalence of acidosis (pH <7.35) at the time of admission. A significantly higher proportion of pneumonia patients were ventilated (p < 0.001); there were no significant differences between the 2 groups in management strategy for those who were ventilated. No significant difference was observed in terms of respiratory-team involvement in their care. A higher proportion of patients with pneumonia had their ‘do not resuscitate’ (DNR) orders issued within 24 h of admission.

Pneumonia per se was associated with poor outcome (table 3), in terms of inpatient and 90-day mortality, increased length of stay and being discharged on oxygen. There were minimal differences with regard to being accepted by early discharge schemes, or in being re-admitted within 90 days of the audit index admission. Progressive adjustment (table 4) for age, sex and other variables that were significantly different between the groups including residence, disability, co-morbidity, albumin and urea levels gradually attenuated the association of pneumonia with mortality and to a lesser extent with a length of stay of at least 7 days. Figure 1 shows the survival comparison between the 2 groups for up to 90 days after admission.

Our study provides a useful insight into the prognosis of COPD exacerbations which are associated with pneumonia compared to non-pneumonic COPD exacerbations.

To be included in this audit there needed to be a clinician-made diagnosis of COPD exacerbation and that would normally have taken into account the CXR appearance in the making of the primary diagnosis.

We found that COPD exacerbations associated with radiological pneumonia have worse outcomes in terms of inpatient mortality and length of stay after taking into account potential confounding factors. The major strengths of the study are: the large sample size, prospective identification of COPD exacerbations, radiological evidence of pneumonia in the presence of pneumonia symptoms, exclusion of suspicious or confirmed malignant cases, a sample distribution across all ages and an ability to take account of case mix and individual prognostic indicators including social factors. Furthermore, this study provides a useful insight in a UK setting into the management of COPD exacerbations when someone presents with pneumonia compared to COPD exacerbation without pneumonia.

In this study, we reported the characteristics of COPD exacerbations which are associated with radiological pneumonia compared to other non-pneumonic exacerbations. This approach is in contrast to that of Pifarre et al. [9], who studied pneumonia patients to gain insight into the characteristics of CAP in patients with COPD. To date, little data is available on the management and course of COPD patients with exacerbations who present acutely with or without pneumonia. As a nationwide evaluation audit project, we were able to gain deeper insight into the management and course of the disease comparison between these 2 groups of patients. There also appeared to be some differences, although minor and perhaps with good reason, between the management of those with pneumonia and those without. Although we found statistical differences between the 2 groups in the clinical course and management, the material differences were not huge apart from a significantly (p < 0.001) higher proportion of patients with pneumonia who were ventilated (16 vs. 12%) and those who had DNR orders within the first 24 h of admission (19 vs. 13%). The DNR orders were more likely to be in place within 24 h of hospital admission for a higher proportion of pneumonia patients. Therefore, it may be interpreted that less intensive therapy is intended in COPD patients with pneumonia compared to those with non-pneumonic exacerbation. However, the fact that a higher proportion of patients with pneumonia also received ventilation (i.e. active intensive management) perhaps indicates that management plans were tailored to the individual’s need rather than to systematic differential treatment. Furthermore, this finding could merely be a reflection of the severity of the disease.

Patients with pneumonia had poorer outcomes, both for mortality and a prolonged length of stay (longer than 7 days) when compared to patients with no evidence of pneumonia in COPD exacerbation. It has been reported that COPD patients hospitalised with CAP had higher 30- and 90-day mortality than patients without COPD [10]. However, the evidence is conflicting, with Snijders et al. [7] more recently reporting no differences in 30-day mortality. Dewan et al. [11] reported that patient host factors may determine treatment outcome in acute exacerbation of COPD. They found that host factors such as FEV₁ <35%, the use of home oxygen, frequency of exacerbation, a history of previous pneumonia and sinusitis, and the use of maintenance steroids were independently associated with treatment failure (defined as a return visit for persistent respiratory symptoms that required a change of antibiotic in <4 weeks). However, they did not specifically examine differences in COPD exacerbation with or without pneumonia.

More recently, it was reported that COPD exacerbations were associated with impaired lung function and increased sputum neutrophilia. Papi et al. [12] reported that a high proportion of patients in their small series (78% of 64 patients included in the study) were associated with viral and/or bacterial infections. Again, they did not differentiate between infective exacerbation due to lower respiratory tract infection and pneumonia. They concluded that respiratory infections are associated with most COPD exacerbations and the severity thereof, especially those with viral/bacterial co-infection.

Holguin et al. [13], in a nationally representative sample of hospitalisations in the USA between 1979 and 2001, reported that any mention of COPD in the discharge diagnosis is associated with higher hospitalisation prevalence and in-hospital mortality from other co-morbidities, including pneumonia. In contrast to our study, their study focused on the presence of COPD as a co-morbidity in patients with other conditions. In the UK, we found that patients with COPD who presented with pneumonia had higher numbers of co-morbidities and also had a higher mortality rate. We were able to examine the extent to which co-morbidity can explain differences in mortality between COPD patients with and without pneumonia (table 4). Interestingly, adjusting for the number of co-morbidities did not attenuate the results further, after adjusting for disability and place of residence. It may be that people with more co-morbidity are likely to have greater disability and are more likely to live in institutional settings.

Our study has limitations. The pneumonia diagnosis was based on radiological appearance rather than on the aetiological diagnosis. However, previous attempts to examine the aetiological diagnosis of pneumonia in COPD patients had several limitations. In a sample of COPD patients (n = 124), the aetiological diagnosis was achieved only in 59% of cases based on valid techniques to identify aetiological diagnosis [14]. One of the limitations of our study is the possible misclassification of some pneumonia cases as non-pneumonia (e.g. atypical) where there may be a mismatch between clinical presentation and radiology. Nevertheless, it is more likely to be the other way round, i.e. the presence of radiological shadow in the absence of clinical signs and symptoms. Furthermore, misclassification of a few cases of radiologically negative pneumonia (i.e. no CXR shadow) is very unlikely to influence the overall outcome in a study of this size; Melbye et al. [15] reported that while there are some patients where CXR may be negative, their findings supported the view that patients with radiological pneumonia as a rule present more serious manifestations of lower respiratory tract pathology than patients with a normal radiograph. Furthermore, CXR is regarded as a reference standard for pneumonia [16], due to multiple factors including a lack of reliability among physicians to elicit physical signs in chest examinations [17], and the over-diagnosis of pneumonia based on chest signs alone [18].

There has been some concern about the increased risk of non-fatal pneumonia with high-dose inhaled steroids in COPD [19]. However, the clinical trial evidence is weak, mainly due to the lack of definition of pneumonia being reported as adverse events in a trial setting, and thus the risk involved is thought to be minimal, judged by the National Institute of Health and Clinical Excellence (NICE) Guideline Development Group expert panel [19]. Our findings suggest that the older people with COPD who have multiple co-morbidities and poor performance status appear to be at higher risk of pneumonia, so clinicians should be aware of this when making decisions about initiating long-term high-dose steroids in this age group. There appears to be a consistent finding that patients with pneumonia in COPD have more co-morbid conditions and are more likely to die or have longer stays in hospital, which is another objective valid outcome measure used by NICE in the UK.

In summary, we found an additional risk for COPD exacerbation outcome in terms of mortality, length of stay >7 days and a 90-day readmission rate in those with a radiological finding of pneumonia. These patients also received antibiotics and ventilation significantly more often. The results are clinically relevant, according to them treatment standards such as CXR for hospital-admitted COPD exacerbation should be obligatory in order to identify patients with radiological pneumonia and to commence appropriate treatment as soon as possible. While severity assessment tools for pneumonia exist, less is researched in the area of prognostic rules in COPD with pneumonia. Risk stratification of COPD patients with pneumonia requires further evaluation in future studies.