Factors Associated With Outcomes of Acute Exacerbations of Chronic Obstructive Pulmonary Disease

Abstract

The purpose of this article is to provide a general review of the current literature on the factors associated with the outcomes of hospitalizations, survival and health‐related quality of life in acute exacerbations of chronic obstructive pulmonary disease (AECOPD), highlighting the limitations and the complexities in interpretation of the results of current studies. There is no consensus definition for AECOPD; onsets may be difficult to define and the determination of duration elusive. The prevalence of acute exacerbations of COPD (AECOPD) in the community appears to be underestimated as exacerbations are underreported by patients and their doctors. Hospitalization for COPD is due mainly to severe AECOPDs which drive the cost of care. There are few longitudinal epidemiological studies on factors associated with hospitalizations for AECOPD. The results of current studies do not allow clear differentiation between associations that are predictors of event, the consequences of the event, or indicators of severity. Strategies to reduce severe exacerbations of COPD include pharmacological treatment, vaccinations, pulmonary rehabilitation, and home care programs. The optimal strategy for the reduction of hospitalization in COPD remains unclear. Long-term interventional studies are needed to provide clearer information for the prevention of exacerbations and hospitalizations in COPD.

• Chronic obstructive pulmonary disease
• AECOPD
• Acute exacerbations
• Outcomes

Introduction

Chronic obstructive pulmonary disease (COPD) remains an under‐diagnosed, poorly understood and inadequately managed medical care problem today [[1]][[2]]. The outcomes of care remain poor, with particularly high rates of hospital readmissions for acute exacerbations of COPD (AECOP) [[2]]. Over half of COPD patients who are hospitalized for AECOPD are expected to be readmitted at least once in the ensuing 6 months [[3]][[4]][[5]][[6]] and nearly a fifth of patients discharged from emergency departments require hospitalization [[7]]. Hospitalization for AECOPD accounts for a large part of the high healthcare expenditure for COPD, estimated to be 40% of the total direct cost of medical care for COPD in the U.S. [[4]].

Exacerbations of chronic obstructive pulmonary disease are important events in the progression of COPD because they are the main cause of hospitalizations for patients with COPD, and are linked to poor outcomes of deaths [[5]][[8]][[9]][[10]][[11]], further deterioration of lung function [[12]] and reduced quality of life [[13]] and account for most of the direct cost of the disease [[14]][[15]][[16]][[17]][[18]][[19]][[20]][[21]]. Exacerbations are described as the “major battlefront of the physician's war on the disease” [[22]]. The prevention and amelioration of exacerbations have become an important primary outcome measure in the evaluation of effectiveness in drug trials in COPD [[23]][[24]][[25]][[26]][[27]]. Many putative risk factors for exacerbations have been proposed [[28]][[29]][[30]][[31]][[32]][[33]] and many are potentially modifiable [[31]], presenting a key strategy for reducing the cost of healthcare for COPD.

The purpose of this article is to provide a general review of the current literature on the factors associated with hospitalizations, and outcomes of survival and health‐related quality of life in acute exacerbations of chronic obstructive pulmonary disease (AECOPD), highlighting the limitations and the complexities in interpretation of the results of current studies.

The Definition of Exacerbation and Severity

Acute exacerbations of COPD (AECOPD) are a common cause of morbidity and mortality. There is no consensus definition for AECOPD. A clear and standardized definition of exacerbation is needed for an accurate and prompt diagnosis and hence treatment of exacerbations and the identification of patients at special risk for recurrent exacerbations. Such a definition would also facilitate patient education and awareness, disease recognition, research and clinical trials.

Clinical Presentation

Presentations of exacerbation may vary from the familiar symptoms of increases in dyspnoea, in coughing, in sputum production and purulence, and wheezing to the nonspecific symptoms of malaise, fatigue, reduced effort tolerance, insomnia or sleepiness, depression and confusion [[34]][[35]].

The clinical assessment of severity of an acute exacerbations has long been based on a combination of medical history of the patient before the exacerbation, presenting symptoms, physical signs of cardiorespiratory distress, changes in the mental status, lung function tests, arterial blood gas measurements [[34]]. Signs of a severe exacerbation include the use of accessory muscle, paradoxical chest wall movements, worsening or new onset of cyanosis, development of peripheral edema, hemodynamic instability, signs of right hear failure and reduced alertness (Table 1).

Table 1.  Clinical Presentation of Acute Exacerbation of COPD

Medical history and signs of severity of exacerbations of COPD
Medical history	Signs of severity
• Duration of worsening or new symptoms.	• Use of accessory respiratory muscles.
• Number of previous episodes (exacerbations/hospitalizations).	• Paradoxical chest wall movements.
• Present treatment regimen.	• Worsening or new onset central cyanosis.
	• Development of peripheral edema.
	• Hemodynamic instability.
	• Signs of right heart failure.
	• Reduced alertness.

The need for hospitalization for an AECOPD is largely determined by the development of acute respiratory acidosis, the presence of significant comorbidities and the need for ventilatory support (Table 2).

Table 2.  Criteria for Hospitalization for a Patient with an AECOPD

Indications for hospital assessment or admission for exacerbations of COPD
• Marked increase intensity of symptoms, such as sudden development of resting dyspnea.
• Severe background COPD.
• Onset of new physical signs (e.g., cyanosis, peripheral edema).
• Failure of exacerbation to respond to initial medical management.
• Significant comorbidities.
• Newly occurring arrhythmias.
• Diagnostic uncertainty.
⩽̸ Older age.
• Insufficient home support.

*Local resources need to be considered.

Some patients deteriorate further and may need some form of ventilatory support which includes either noninvasive mechanical ventilation using negative or positive pressure devices in the setting of an intermediate or special respiratory care unit or in an intensive care unit [ICU] for invasive mechanical ventilation by endotracheal intubation.

Consensus indications for NIPPV and invasive mechanical ventilation are shown in (Tables 3 and 4).

Table 3.  GOLD Consensus of ICU Admission for AECOPD

Indications for ICU admission of patients with exacerbations of COPD
• Severe dyspnea that responds inadequately to initial emergency therapy.
• Confusion, lethargy, coma.
• Persistent or worsening hypoxemia (PaO2 < 5.3 kPa, 40 mm Hg), and/or severe/worsening hypercapnia (PaCO2 < 8.0 kPa, 60 mm Hg), and/or severe/worsening respiratory acidosis (pH < 7.25) despite supplemental oxygen and NIPPV.

*Local resources need to be considered. [From Ref. [[34]].]

Table 4.  [GOLD] Consensus for Mechanical Ventilation in Exacerbation of COPD

Indications for invasive mechanical ventilation
• Severe dyspnea with use of accessory muscles and paradoxical abdominal motion.
• Respiratory frequency > 35 breaths per minute.
• Life‐threatening hypoxemia (PaO2 < 5.3 kPa, 40 mm Hg or PaO2/FiO2 < 200 mm Hg).
• Severe acidosis (pH < 7.25) and hypercapnia (PaCO2 < 8.0 kPa, 60 mm Hg).
• Respiratory arrest.
• Somnolence, impaired mental status.
• Cardiovascular complications (hypotension, shock, heart failure).
• Other complications (metabolic abnormalities, sepsis, pneumonia, pulmonary embolism, barotraumas, massive pleural effusion).
• NIPPV failure (or exclusion criteria).

FiO₂: Fractional concentration of oxygen in dry inspired gas.

Definition for Research

For many years, the best known definition of an exacerbation was one that had been developed specifically for studies of antibiotics in bacterial exacerbations [[1]][[36]][[37]]. The authors defined exacerbations in terms of symptoms of increased dyspnea, sputum production, and sputum purulence: signs indicative of an infectious etiology. Three practical subtypes were defined, with antibiotic treatment being most useful in type 1 and less so for type 2. Type 1 was defined as increased dyspnoea, sputum volume and sputum purulence; type 2 exacerbations were defined as occurring when two of three symptoms were present. Type 3 exacerbations were defined as occurring when one of three symptoms was present in addition to at least one other respiratory symptoms (as shown in Table 5). However, exacerbations can be provoked by non‐infective causes which include acute air pollution [[38]][[39]], cold weather [[40]], and interruption of regular treatment [[41]]; and the symptoms of exacerbation form a spectrum. The use of such a specific definition would then result in under‐recognition of exacerbations, a finding well documented in a later prospective study of exacerbation of COPD [[13]].

Table 5.  Definitions of Chronic Obstructive Pulmonary Disease Exacerbations Developed for Studies of Antibiotics

Type	Symptoms	Additional minor symptoms
1	An increase in sputum volume
	An increase in sputum purulence
	An increase in dyspnoea
2	Any two from
	An increase in sputum volume
	An increase in sputum purulence
	An increase in dyspnoea
3	A single symptom from	Sore throat or nasal discharge within past 5 days
	An increase in sputum volume	Fever without other cause
	An increase in sputum purulence	Increased wheezing
	An increase in dyspnoea	Increased cough
	Plus at least one minor symptom	Increased respiratory rate > 20% above baseline
		Increased heart rate > 20% above baseline

The best‐studied definition for COPD exacerbation [AECOPD] is that developed specifically for studies of antibiotics in bacterial exacerbations. There are currently no consensus definition for AECOPD. [From Ref. [[42]].]

Attempts at Consensus Definition

A similar but expanded version of the earlier definition [[36]] was described in the British Thoracic Society guidelines [[1]] in which symptoms indicative of an exacerbation include increased sputum purulence and/or volume, increased dyspnoea and/or wheeze, chest tightness, and fluid retention.

In 1999, in one of the first attempt at a wider consensus, a transatlantic group of respiratory specialists proposed an operational consensus definition [[35]] for further development towards a globally accepted definition. An exacerbation was described as “a sustained worsening of the patient's condition, from the stable state and beyond normal day‐to‐day variations, that is acute in onset and necessitates a change in regular medication in a patient with underlying COPD.” A simple practical severity sub‐classification into mild, moderate and severe, based on the extent of health care utilization [home care, need for medical assistance, need for hospitalization] was suggested, modified further by additional factors such as the underlying severity of COPD, the number of exacerbations per year and the presence of underlying comorbidity (Table 6).

Table 6.  Staging of a COPD Exacerbation Based on Health‐Care Utilization

Severity	Level of health‐care utilization
Mild	Patient has an increased need for medication, which he/she can manage in own normal environment
Moderate	Patient has an increased need for medication and feels the need to seek additional medical assistance
Severe	Patient/caregiver recognizes obvious and/or rapid deterioration in condition, requiring hospitalization

It is also crucial to note the severity of the underlying COPD, any Comorbid conditions, and the frequency of exacerbations.

[From Ref. [[35]].]

Further specific refinements of this broad classification was proposed by S. Burge and J.A. Wedzicha [[42]]. They suggested definitions of different complexities for different applications. For example, a specific definition for the study of a bacteriological cause; a simple definition for the retrospective diagnosis from data base source; a prospective definition based on changes from an agreed baseline from entries in patient diary cards. The authors also suggested that the severity scale proposed by the translantic consensus group [[35]] could be further developed to be more comprehensive and to incorporate exacerbations management in hospital and at home, and the need for parenteral corticosteroid and antibiotic and the severity of arterial blood gas abnormalities (Table 7). Exacerbation duration determined from diary cards could also be included as severity marker. The development of this definition continues.

Table 7.  A Scale for Exacerbation Severity Incorporating Exacerbations Managed at Home and in Hospital

Mild	An exacerbation treated with antibiotics but no systemic corticosteroid. If no blood gases are available the absence of respiratory failure is assumed.
Moderate	An exacerbation treated with parenteral corticosteroids with or without an antibiotic. If no blood gases are available the absence of respiratory failure is assumed.
Severe	Type 1 respiratory failure with hypoxaemia but no carbon dioxide retention or acidosis; Pa,O2 < 8 kPa (60 mmHg) and Pa,CO2 < 6 kPa (45 mmHg)
Very severe	Type 2 respiratory failure, compensated with hypoxia, carbon dioxide, retention but no acidosis; Pa,O2 < 8 kPa (60 mmHg), Pa,CO2 < 6 kPa (45 mmHg) and hydrogen ion concentration < 44 nM (pH > 7.35)
Life‐threatening	Type 2 respiratory failure, decompensated with acidosis and carbon dioxide retention; Pa,O2 < 6 kPa (45 mmHg) and hydrogen ion concentration > 44 nM (pH < 7.35)

Pa,O₂: arterial oxygen tension: Pa,CO₂: arterial carbon dioxide tension.

[From Ref. [[42]].]

The importance or lack of importance of “under‐recognition” and “unreported exacerbations” remains controversial. It could be argued that if patients do not report a “change in symptoms” that may require an alteration in available medications such as an increase in the frequency of an inhaled short‐acting bronchodilator, without additional adverse outcomes, medical interventions nor medical visits, then this sequence of events may be unimportant. Whether such events should be classified as “exacerbations” or merely diary card recordings of events outside the normal day‐to‐day variation in symptoms remain unresolved. The use of such definitions may impact on the interpretation of clinical drug trials.

Clearly, there is a need for a more comprehensive and universally applicable definition in clinical practice, in clinical studies, in drug trials, and in health cost evaluation.

Determination of Exacerbation Frequency

Exacerbations are widely believed to be infrequent during the early phase of COPD and are largely a problem of moderate to severe COPD [[1]][[34]] (Figure 1). It is possible that such exacerbations are likely to be unrecognized since COPD is rarely identified in patients early in the course of the disease. Exacerbations of moderate to severe COPD are also estimated to be substantially underreported by the patients, 50% in one study [[12]]. There are several problems confronting the determination of exacerbation frequency reported in studies. 1) The lack of physiological changes such as FEV1 in any of the definitions; 2) Exacerbation frequency varies seasonally and can be triggered by infection [mainly bacterial and viruses], and by non‐infective factors such as smoking, air pollution and non‐adherence to medication; 3) The lack of biomarkers of exacerbations; 4) Different definitions for drug trials; and 5) Different data collection sources: primary care records, emergency or hospitalization records, retrospective database, or prospective diary card record. Retrospective determination can be made from information from health care database; but these may not be able define the duration of an exacerbation as it could not discriminate between a new exacerbation and a relapse. Prospective determination would require time consuming and expensive daily diary card monitoring but is the most accurate way of determining the frequency [[12]][[29]] and duration of an exacerbation [[12]][[43]].

Figure 1. Incidence of exacerbations per year/mean related to forced expiratory volume in one second (FEV1) in the placebo arms of three 3‐yr studies. [Reproduced from Ref. [[42]].] The open bars are from the inhaled steroid in obstructive lung disease in Europe [ISOLDE] study [From Ref. [[49]]]; the hatched bar is from the Copenhagen City lung study [From Ref. [[45]]] and the shaded bar is from the European respiratory society study on chronic obstructive pulmonary disease. [From Ref. [[46]].]

The use of patient‐directed daily diary cards recording was formalized by Seemungal and colleagues who used it to detect exacerbations prospectively and monitor the frequency in a study of the time course and recovery of exacerbations of COPD [[12]]. Respiratory symptoms were first classified as “major” symptoms (dyspnoea, sputum purulence, sputum amount) or “minor” symptoms (wheeze, sore throat, cough, and symptoms of a common cold, namely nasal discharge or congestion). An exacerbation was then defined as an increase for two consecutive days of 2 or more of the 3 majors or 1 major and 2 minors according to criteria modified from the stratification criteria by Anthonisen and colleagues [[36]]. Such patient‐directed diary cards had similarly been employed to evaluate the time course of an exacerbation after it had begun [[43]].

Determination of Time Course of Exacerbation

Information on the duration and recovery of an exacerbation is based on a limited number of studies [[12]][[43]]. The duration depends on the measurement [symptoms, sputum, or health status] employed, the severity of preexisting COPD and the patient population studied: primary care or hospitalized patients.

In the first comprehensive study on the changes in lung function and symptoms before and after exacerbations, Seemungal and colleagues found that worsening symptoms but not lung function occurred before the onset of an exacerbation. Furthermore, symptoms and lung function changes usually improve within a median 7 days but full recovery may be delayed beyond 3 months in about 7% of exacerbations (Table 8).

Table 8.  Recovery from Exacerbation in PEFR and Total Symptom Score in 91 Patients with 504 Exacerbations

	PEFR (IQR)	Symptoms (IQR)
Median time to recovery, d	6 (1 to 14)	7(4 to 14)
% Exacerbations recovering within 35 d	75.2	86.1
% Exacerbations recovering within 91 d	80.2	90.9
% Exacerbations in which the next exacerbation occurs before complete recovery in PEFR	3.4	1.4
% Exacerbations with indeterminate recovery	9.3	3.1
% Exacerbations that do not recover at 91 d	7.1	4.6

IQR = interquartile range.

*Results are presented as median (IQR).

^†Recovery could not be determined for these exacerbations because of missing data.

[From Ref. [[12]].]

Determination of Exacerbation Severity

In clinical practice, the severity of an exacerbation is descriptive, based on a generalized assessment of the patient's medical history before the exacerbation, symptoms, physical examination, lung function, arterial blood gas measurements and other laboratory tests [[34]]. In clinical research, there is as of yet no consensus definition [[35]][[42]].

Exacerbation and Hospitalization

The indications for hospital admission depend not only on the clinical severity of the acute exacerbations of COPD but also on other factors such as a lack of social support and the presence of comorbidities [[34]]. In general, COPD‐related hospitalizations reflect the magnitude of severe exacerbations in a community. However, the proportion of exacerbations resulting in hospitalizations may vary depending on the patterns of health care utilization of the country [[14]].

The Burden of Exacerbation and Hospitalization

The health burden of exacerbations and the economic burden are closely linked. Studies have shown that 80% of patients with moderate to severe COPD experience at least one exacerbation within a 3‐year period [[44]][[45]][[46]]. Despite the considerable burden, exacerbations are under‐recognized, under‐reported and under‐treated [[14]].

Economic Burden

It is widely recognized that a strong correlation exists between the stage of the disease and the treatment costs, due to increased utilization of clinics, emergency room visits, long‐term oxygen therapy and hospitalization being associated with advanced disease [[34]]. Studies in the US have shown that patients with the most severe disease which form a small proportion of all COPD patients (about 10%) account disproportionately for half of all the expenditure, due largely to hospitalizations [[15]][[20]][[21]]. Hospitalizations are responsible for a substantial portion of the direct medical cost of COPD which is estimated to be twice that for asthma [[15]].

Exacerbations of COPD are the key reason for greatly increased use of secondary health care resources, such as ER visits, and hospitalizations. The results of a large‐scale seven country survey in Europe and North America demonstrate that the majority (52–84%) of all direct costs associated with COPD were due to hospitalizations [[14]]. Therefore any measure aimed at preventing and treating exacerbations could significantly reduce the overall economic cost of the disease.

Exacerbations and Health Care Utilization

In most countries, exacerbations are closely associated with increased ER visits and hospitalization and their associated direct and indirect costs [[14]]. However, the economic impact may not be the same as has been shown by a large comparative international survey of the burden of COPD in countries in Europe and North America. In 5 of the 7 countries which took part in the survey, exacerbations parallel hospitalizations. However, there were between‐country differences in hospitalization rates and cost patterns, which could not be explained by the patient characteristics alone and may be associated with differences in the pattern of health care utilization and management approaches (Fig. 2). An inverse relationship was found between the rate of hospitalization and the rate of use of long acting beta agonists [[14]], suggesting that the rates of hospitalization is modifiable by specific interventions.

Figure 2. International comparison of the burden of health care utilization due to COPD. [From Ref. [[14]].]

The Risk/Associated Factors of Exacerbations

Exacerbation of COPD is a driver of the disease and there is a need for better understanding of factors associated with its pathogenesis. Exacerbation onset, frequency and severity are difficult to establish [[42]] but they are important as a growing number of prospective studies and randomized trials are using exacerbation onset as a point of recruitment, and exacerbation frequencies and severity as special outcome measurements for interventional studies.

The factors associated with acute exacerbations and hospitalization in COPD patients are poorly understood (Table 9). The potential relationships between predictors of hospitalization, predictors of outcome after hospitalization and predictors of mortality are unclear. A range of potential factors have been studied both as risk factors of exacerbations resulting in hospitalization as well as predictors of mortality and other outcomes. Furthermore, the number of studies that examined these factors for hospital admissions for AECOPD are limited and at times conflicting.

Table 9.  Risk Factors of Exacerbation/Hospitalization

Risk factors for potential intervention in exacerbation/hospitalization for COPD
Modification—unclear	Modification—potential
• FEV1	• Infection‐ bacteria/ virus
• Arterial blood gases	• Reduced physical condition
• Basal metabolic rate	• Dyspnoea
• Disease progression	• Smoking
	• Chronic mucous hypersecretion
	• Comorbidity
	• Psychological factors
	• Health status [quality of life]
	• Withholding of medications

The proposed risk or protective factors for exacerbations resulting in hospitalizations include forced expiratory volume in one second (FEV1) [[6]][[28]][[29]][[30]][[32]][[33]], blood gases [[28]][[32]][[33]], co‐morbidity [[30]][[32]][[33]], chronic mucus hypersecretion [[30]][[47]], muscle weakness [[48]][[49]], poor nutritional status and low body mass index [BMI] [[28]][[30]][[32]][[33]], socio‐economic status and support [[32]], number of previous physician visits or hospital admissions [[30]][[32]][[33]], QOL scores [[50]], influenza vaccination [[30]][[32]][[33]], pneumococcal vaccination [[32]][[33]], pulmonary rehabilitation [[32]][[33]], inhaled corticosteroids and long‐term oxygen therapy [[32]][[33]].

The results for these studies of acute exacerbations and hospitalization are inconsistent for some factors such as FEV1 [[6]][[28]][[29]][[30]][[32]][[33]], BMI [[28]][[30]][[32]], and smoking [[28]][[32]] and perplexing for others, such as for influenza vaccination [[32]] and care by a pulmonologist [[32]]. These studies employed different designs and different populations of patients in different settings (primary care, hospital, intensive care unit), with different selection and matching criteria, and most studies investigated a limited number of risk factors. Together, these may partly account for the inconsistencies of results. Two larger multi‐site studies [[30]][[31]][[32]][[33]] have attempted to investigate a wide range of potential risk factors in one study, which could fully take into account potentially unknown confounding among variables. The multi‐centre studies have added to the smaller studies in various ways: by studying a comprehensive list of potential risk factors, by employing case control or prospective longitudinal designs, and by using multivariate analysis in an attempt to eliminate confounders, but they could not resolve the uncertainty of cause and effect and thus the importance and relative role of these putative risk factors. Longitudinal randomized controlled interventional studies are thus needed. Other potential risk or protective factors such as psychological well‐being, treatment adherence and social support have not been specifically studied.

Some of these predictive factors for hospitalized exacerbations have also been studied as potential predictors of outcomes or health care utilization or as a measured outcome of exacerbation: FEV1 and treatment outcome [[51]]; chronic hypersecretion and death in infective exacerbation [[52]]; low BMI and readmissions [[49]]; QOL scores and mortality [[6]][[13]].

Thus, studies of acute exacerbations fall under two main categories: studies of risk factors of exacerbations and hospitalizations and studies of the resultant outcomes such as mortality, length of hospitalization and readmissions. There is much overlap because of the common factors that influence exacerbations and hospitalizations and associated outcomes and the difficulty in distinguishing true independent risk factors or predicators of exacerbations from markers of severity, or even predictors of outcomes of exacerbations in observational studies.

The factors associated with exacerbations of COPD can be studied by recruitment in the community [[12]][[53]][[54]] or in the hospital [[5]][[32]][[33]]. Studies of community based exacerbations can be extrapolated to the population, but hospital based exacerbations are more accessible for studies and are clinically and economically important. Hospitalizations are finite end points and are used in the economic evaluation of severe exacerbations which are difficult to identify in the community [[15]]. Hence, hospitalizations, and in particular recurrent hospitalizations or readmissions, are important outcome measures in studies of risk factors of exacerbations and their intervention.

Prevalence of Modifiable Factors

A high prevalence of potentially modifiable risk factors had been reported in several studies [[31]][[55]][[56]][[57]]. Easily identifiable and modifiable risk factors in hospital based populations include the unmet aspects of stable COPD management such as a lack of influenza vaccination, pulmonary rehabilitation, and long‐term oxygen therapy [LTOT] non compliance, faulty inhaler technique, and current smoking.

The overall influenza vaccination prevalence rates for patients admitted with AECOPD varied from between 72% [[31]][[55]] and 40% [[56]]. Pulmonary rehabilitation rates are low, and contrary to the guideline, recommendations based on the many benefits of pulmonary rehabilitation in survival, health related quality of life, and exercise capacity [[34]][[37]]. Reported pulmonary rehabilitation rates vary from 5–8% in Switzerland [[57]] to 14% in Barcelona [[31]] to a high in New Zealand 25% [[56]].

Non‐prescription of LTOT according to established guideline recommendations [[37]][[58]] is high (28%) in patients hospitalized for COPD exacerbation and could be improved. LTOT non compliance rates defined as LTOT < 15 hr/day vary in different studies, from 20% to 55% [[59]][[60]].

Current smoking prevalence rates vary but remain substantial: 26% [[31]] and 33% [[56]] and 34% [[28]]. Exposure to smoking at home [passive smoking] is high at 21%. Although the role of passive smoking has not been evaluated [[61]], it is likely to be harmful, leading to accelerated decline in lung function and as a trigger for AECOPD.

Interpretation of Risk Factors

Many such risk factors relating to patient and disease characteristics have been proposed but there are few systematic studies and results are often conflicting. These inconsistent results may be due to the fact that the studies employed: 1) different designs: retrospective audit data [[62]]; prospective cohort [[33]]; secondary analysis of a RCT [[54]][[63]]; 2) different entry criteria [[28]][[29]]; 3) different definition of an exacerbations [[12]][[50]][[62]]; 4) different populations [[12]][[29]][[50]][[53]][[63]]; 5) different timing of recruitment—during stability [[12]][[28]][[29]][[63]] and during hospitalization for COPD exacerbations [[31]][[54]]

In such studies, statistical associations of a number of variables recorded at the time of the exacerbation or hospitalization or after the event [[13]] are interpreted as predictors of the event. There are alternative explanations of this assumption. Some of these associations may be mere accompaniments or are “confounders by indication” [[33]][[64]] rather than predictors of the event or outcome. It is also conceivable that in some patients, frequent admissions may have resulted in, rather than resulted from, the increased use of sedatives, anticholinergic, oral steroids, lower health‐related quality of life scores, and a lower FEV1. Similarly, the frequent admissions could have resulted in malnutrition, depression, and the lack of caregiver support from caregiver fatigue. These circuitous links are well illustrated by studies which showed that FEV1 [[51]] and health‐related quality of life [[13]] may be predictors of exacerbations and yet they in turn may result from frequent exacerbations [[13]][[63]]. Despite the confusion in the literature, studies of risk factors of exacerbations and hospitalizations are important and form an essential preliminary systematic approach for generating hypotheses of potential risk factors of various outcomes to be tested in further interventional studies.

Infection

About two thirds of exacerbations of COPD are attributed to infection of the tracheobronchial tree [[54]][[65]][[66]][[67]][[68]] and air pollution [[39]], while no cause can identify up to one‐third of cases [[5]][[69]]. The most common pathogens consistently isolated from patients with COPD are Haemophilus influenza, Moraxella catarrhalis, and Streptococcus pneumoniae. There is consistent evidence that viruses are likely to account for one‐third of all cases [[36]][[70]], while serological evidence indicates that Mycoplasma pneumoniae and Chlamydia pnuemoniae accounts for 5 to 10% of exacerbations [[71]][[72]].

The Role of Bacteria

The role of bacterial infections, once believed to be the main cause of exacerbations, is controversial [[68]][[73]][[74]][[75]][[76]]. Studies [[77]][[78]][[79]] have found that patients with COPD often have pathogens in their airways in the absence of exacerbations while some exacerbations occur in the absence of bacteria [[71]], leading to the conclusion that bacteria colonized airways rather than caused infection [[80]][[81]]. There is better consensus on the spectrum of bacteria [[66]][[67]][[72]][[82]]. A recent prospective study [[69]][[83]] has shed some light on the bacterial role of exacerbation. The authors found that the isolation of new strains of H. influenzae, M. catarrhalis, and S. pneumoniae in patients with COPD was more strongly associated with acute exacerbations than those without new strains. The unifying theory is that as the airways of COPD patients are often colonized by bacteria, an emergence of new strains may be the underlying mechanism/risk factor for triggering a new exacerbation of COPD (Table 10).

Table 10.  Greater the Prevalence of New Strains Greater the Relative Risk of Exacerbation of COPD

Relative Risk of an Exacerbation According to Whether a New Strain of Bacterial Pathogen was Isolated
New strain	Frequency of exacerbation		P value	Relative risk (95% CI)
	New strain	No new strain
	No. of exacerbations/total no. of visits (%)
Any strain	89/270 (33.0)	213/1385 (15.4)	< 0.001	2.15 (1.83–2.53)
Haemophilus influenzae	38/145 (26.2)	257/1503 (17.1)	< 0.001	1.69 (1.37–2.09)
Moraxella catarrhalis	41/84 (48.8)	261/1571 (16.6)	< 0.001	2.96 (2.39–3.67)
Streptococcus pneumoniae	8/25 (32.0)	294/1630 (18.0)	0.01	1.77 (1.14–2.75)
Pseudomonas aeruginosa	3/22 (13.6)	297/1631 (18.2)	0.38	0.61 (0.21–1.82)

*The relative risk of an exacerbation was for the presence of a new strain in sputum, as compared with its absence. Relative risks were calculated with the use of generalized estimating equations, CI denotes confidence interval.

^†Seven visits were excluded because of simultaneous isolation of new strains of M. catarrhalis (six visits) and P. aeruginosa (one visit).

^‡Two visits were excluded because of simultaneous isolation of new strains of M. catarrhalis.

(From Ref. [[69]].)

The Role of Viruses

The role of viruses in exacerbations of COPD has not been adequately studied partly due to the limitations of standard diagnostic methods and partly to the lack of incentive because of the absence of effective prevention or treatment for viruses other than influenza. A limited number of studies [[36]][[70]] showed that viruses were isolated in 1/3 of all exacerbations but there is minimal information on the relative frequencies of the different respiratory viruses in exacerbated and quiescent COPD. The use of molecular techniques could facilitate our understanding of viral epidemiology in COPD by determining the prevalence of the different respiratory viruses.

The Impact of Lower Respiratory Tract Infection on Outcome

For many years, it was unclear whether infective episodes in COPD affect the progression of disease. Early studies from the UK [[84]][[85]] and from Canada [[86]] found no such association while a prospective cohort study of 4 years from the US [[87]] found that a more rapid decline in FEV1 was associated with more frequent episodes of respiratory infection. In the Lung Health Study, mild COPD patients were followed for 5 years in a prospective, randomized study of smoking cessation, and the authors found that lower respiratory tract infection was associated with a decline in FEV1 in smokers only, suggesting an interaction between smoking and infection [[63]].

FEV1

Forced Expiratory volume in the first second (FEV1) is the standard objective measure of airflow obstruction in clinical practice and is a commonly used end‐point in clinical trials because the measurement is reproducible and relatively easy to perform. It is also the defining measurement used in the guideline classification of severity of COPD [[34]][[37]] (Table 11).

Table 11.  Studies on the FEV1 as Associated Factor in Exacerbations in or Hospitalizations for Chronic Obstructive Pulmonary Disease [COPD]

Ref	Year	Size	Subjects	Design	Setting	Statistics	Results	Duration	Outcomes	Country
Decramer [Ref. [[48]]]	1997	57	Severe COPD grouped according to hospitalizations	2 group descriptive comparison	Stable COPD No defined selection criteria	2 group comparison	FEV1 not different between the groups	Cross sectional; descriptive	Health care utilization	Belgium
Collet [Ref. [[29]]]	1997	381	Mild to severe COPD	Randomized double blind control of immunostimulant. Secondary analysis of data	Ambulatory patients	Regression models	More severe FEV1, associated with increased hospitalization	6 months	Acute exacerbations in 6 months hospitalization QOL	Canada
Kessler [Ref. [[28]]]	1999	64	Moderate to severe COPD	Prospective cohort	Stable ambulant COPD	Univariate analysis	FEV1 does not predict hospitalization; PaCO2 & PPa independent predictor,	2.5 years	Time to 1st hosp for exacerbation; clinical & functional ‐ ABG, pulmonary hypertension	France
Niewoehner et al. [Ref. [[51]]]	2000	271	Mod to severe COPD	Randomized control	Hospitalized patients	Univariate & multivaliate analyses	FEV1 & 1st few days predict clinical outcomes	30 days with follow‐up 6 months extension	Primary outcome: 30 day treatment failure	US
Garcia‐Aymerich [Ref. [[32]][[33]]]	a) 2001	172	Mod to severe	a) Case control.	Hospitalized patients	Multivariate analysis	FEV1 independent predictor of admission	1 yr recruitment	Clinical & adherence to medicine, social support, physical activity	Spain
	b) 2003	340		b) Prospective cohort.

Since exacerbations are largely a clinical problem of moderate to severe COPD, it would be reasonable to expect that the FEV1 would be a reliable predictor of exacerbations and hospitalization and correlate well with clinical outcomes. FEV1 has been found in some studies [[29]][[30]][[32]][[33]] to be a strong predictor of hospital admissions for acute exacerbations, but not in others [[6]][[28]][[48]]. A possible reason for the discrepancy is that two of the latter studies [[28]][[48]] included small numbers of patients who were selected on a narrow and low range of FEV1. A positive relationship of FEV1 and risk of hospitalization for AECOPD is all the more coherent given the consistent findings that have been observed of the strong impact of FEV1 in predicting mortality [[47]][[88]][[89]] and poor overall clinical outcomes that included death, readmission, intubation and intensification of drug therapy [[51]].

Dyspnoea

This has not been fully studied. One study has reported that the degree of dyspnoea, in lieu of FEV1, was associated with increased risk of relapse of acute exacerbations following ambulatory treatment for an index acute exacerbation [[30]].

Blood Gases

Earlier studies have demonstrated a positive association between blood gas impairment and the risk of hospitalization for AECOPD [[28]][[48]] while later studies which employed a multivariate analysis to correct for potential confounders failed to find this association [[32]][[33]]. Arterial blood gas data present a problem in interpretation as they could be collected under different conditions (room air or with oxygen therapy), and this could possibly explain the observed lack of association. Alternatively, they may be a marker for disease severity rather than an independent risk factor for hospitalization in AECOPD.

Body Mass Index (BMI)

Two previous studies of small groups of highly selected patients demonstrated that low BMI [[28]][[49]] were related to increased risk of hospital readmission, but other studies that included BMI as a study variable in less selected groups of patients did not demonstrate this association [[30]][[32]]. A possible explanation for a lack of positive association of acute exacerbation with BMI may be that anthropomorphic measures of nutritional status in adults, such as BMI < 20 kg/m2, may not be an adequate indicator of poor nutritional status in different populations of COPD patients. Alternatively, BMI is merely a marker of disease severity and a confounder.

Chronic Mucous Hypersecretion

The role of chronic mucous hypersecretion in association with hospitalization is unresolved. There is some evidence from a large prospective community longitudinal cohort [data from the Copenhagen City Heart study and hospitalization data from the National registry], that exacerbations are more frequent in those with regular sputum production and are associated with both an excess FEV1 decline, death from respiratory failure [[52]][[53]], and an increased risk of subsequent hospitalization in severe COPD [[47]]. However, this is a difficult concept as chronic bronchitis is defined by chronic mucous hypersecretion, a characteristic in the majority of patients with COPD.

Physical Activity

The benefits of pulmonary rehabilitation are well established. Physical conditioning is a component of all pulmonary rehabilitation programs which have been shown to reduce the number of hospitalizations and days in hospital [[34]]. There is also some evidence that the level of usual physical activity of the patient is associated with a reduced risk of hospitalization.

In a prospective study of the risk of rehospitalization, a cohort of 340 patients with a previous hospitalization was followed up for 1 year to assess the association between readmission for a COPD exacerbation and a wide range of modifiable potential risk factors [[33]]. The authors found a 63% readmission rate and a 29% mortality rate during the period of the study and a lower FEV1 and hypoxaemia was associated with increased readmissions while a self reported high level of usual physical activity was independently and strongly associated with a 46% reduction in the risk of readmission due to exacerbation.

In a retrospective audit of 43 hospitals in the UK, C.M. Roberts [[62]] and colleagues found that the premorbid health performance status of the patient is the single most powerful predictor [more than increasing age or worsening lung function] of death in patients admitted for exacerbation of COPD. A bed‐bound patient was 20 times more likely to die than one who had normal activity, with a mortality rate of 38.5% at 3 months. This would seem to justify a change in the routine clinical practice of physicians to focus on this factor for intervention.

Comorbidity

Comorbid conditions are common in COPD. Pneumonia, heart failure, and pulmonary embolism have been listed by the European Respiratory Society as “causes” or “triggers” of exacerbation of COPD [[58]] and are confounding conditions in the assessment of exacerbations of COPD. They can also impact negatively on the relationship between the health‐related quality and COPD [[90]]. The evidence for comorbidity as a risk factor for hospitalization is conflicting. Comorbidity has been identified as a risk factor for hospital admissions in ambulatory patients with mild to moderate COPD in the EOLO Study [[30]]. However, two other studies [[28]][[32]] of patients with moderate to severe COPD found that comorbidities had no significant association with increased risk of hospitalization for an acute exacerbation of COPD. Different patient selection criteria may be a possible explanation for the discrepancy.

Psychological Factors

The prevalence of depression is known to be high in patients with advanced COPD [[91]]. The development of psychological sequelae in conjunction with physical illness is recognized to have an impact on subsequent outcomes of care and survival [[8]][[92]]. Very few studies have explored the role of psychological factors in determining the risk of increased acute exacerbations and hospital admissions. One previous study [[32]] has investigated sedative consumption (as a ‘lifestyle’ variable, indicating psychological disturbance) with no positive association found. Psychotropic drug use may be considered a surrogate marker of more severe clinical depression, and could be used to further study the relationship between depression and frequent hospital readmission for AECOPD.

Age, Gender, Smoking

Neither age nor gender appeared to be associated with frequent hospital admissions [[6]][[28]][[32]]. The role of smoking has only been studied in two previous studies [[28]][[32]]. The first study [[28]] found no significant impact of smoking on the risk of hospitalization; the second study found paradoxically that current smoking, compared with ex‐smoking, was associated with reduced risk of COPD admission [[32]].

Medical Care Factors and Prescriptions

The position of these factors in hospitalization is paradoxical and counter‐intuitive. In some studies of patients who had previous hospitalizations [[32]][[33]] factors related to medical care, such as care undertaken by pulmonologists versus general practitioners, prescriptions [taking inhaled anticholinergic], pulmonary rehabilitation, and pneumococcal and influenza vaccination, have been found to be associated with higher risks of admission for COPD. Caution is needed in the interpretation of these results as the previous admission may “confound by indication” as the type of medical care and prescriptions are more likely to be given to patients perceived to have higher risk of hospitalization [[64]]. Hence, the role of these individual factors in hospitalization can only be adequately evaluated in intervention studies in which other confounding factors are controlled [[42]].

Relative Importance of Various Risk Factors

From the perspective of both the patient and the health care provider, the most important risk factors for exacerbations are those that result in hospitalization. To date, intense research activities have been centered on the use of pharmacological means to modify exacerbations and prevent hospitalization. Surprisingly few interventional studies have been conducted in search of a better understanding of the relative roles of non‐pharmacological factors such as smoking, physical and psychological deconditioning, and viral infection due to viruses other than influenza. Interventional studies of these factors will provide crucial answers to how these triggers cause clinical exacerbations of COPD and hence their prevention.

The Consequences or Outcomes of Exacerbations

Factors associated with exacerbations could be predictors or the consequences of the event. Studies on the impact of exacerbations on the progress of the disease have not been conclusive. The observed associations/consequences of exacerbations include increased mortality, frequent hospitalization and a negative impact on the health‐related QOL. However, it is unclear whether exacerbations of COPD constitute an independent risk factor for the long term accelerated decline in lung function [[63]] or whether recovery to a pre‐exacerbation level is the norm [[93]].

Mortality

COPD is the 4th leading cause of death world wide and the rate is still rising. This marked mortality over time for populations with stable COPD is well illustrated by earlier studies of large cohorts in different countries followed up for lengthy period periods [[94]][[95]][[96]][[97]].

Hospitalizations for acute exacerbation of COPD are important milestones in the life of the COPD patient as they are associated with further increase in morbidity and substantial mortality rates after discharge. The reported mortality in the COPD patients who have hospitalized exacerbations is dependent on the severity of their disease [[10]] and ranges from a 1‐year mortality of 23% for hospitalized patients [[8]][[10]] to 38% at 3 months for bed‐bound patient [[62]] to 59% at 1 year for ICU patients [[9]]. Mortality is lower in studies of outpatients, even when the disease is very severe: a 5‐year 31% mortality for patients of FEV1 < 1000 ml [[96]], and a 3‐year rate of 23% for FEV1 < 35% predicted [[97]] (Table 12).

Table 12.  Mortality Rates in Hospitalized Patients Reported in Different Studies

Reference	Date	Size of Cohort	Country	Mortality Rate	Period of Follow‐up	Setting	Design
Seneff et al. [Ref. [[9]]]	1995	n = 362	US	24%–30% (in hospital) 59% ICU	1 year	ICUs (multi sites)	Prospective cohort
Connors et al. [Ref. [[5]]]	1996	n = 1,016	US	11% in hospital	2 years	Hospitals (multi sites)	Prospective cohort
				43% 1 yr
				49% 2 yr
Roberts et al. [Ref. [[62]]]	2002	n = 1,373	UK	14% (whole group)	3 months	Hospital (multi sites)	Retrospective audit program
				38% (bed bound)
Almagoro [Ref. [[8]]]	2002	n = 135	Spain	22% 1 yr	2 years	Hospital	Prospective cohort
				35.6% 2 yr
Garcia‐Aymerich [Ref. [[33]]]	2003	n = 340	Spain	29%	1 year	Hospital (multi sites)	Prospective cohort
Groenegewen [Ref. [[10]]]	2003	n = 171	Netherlands	23%	1 year	Hospital	Prospective cohort
				35% (ICU)

The predictive factors for increased COPD‐related mortality in general have been identified as low FEV1 [[94]][[97]][[98]] older age [[94]][[97]]; cardiac factors [[94]][[98]] PCO2 [[98]], and low BMI [[99]][[100]][[101]]. The factors associated with mortality for COPD patients after an acute exacerbation included high PaCO2, low oxygen saturation, BMI [[5]][[8]][[10]], older age [[5]][[11]][[102]], cardiac factors [[5]][[11]][[102]], comorbidity [[8]][[102]], marital status [[8]][[103]], severity of illness [[5]], health‐related quality of life [[8]], depression [[8]][[92]], poor functional status [[5]][[8]] and frequent readmissions [[5]].

Frequent Hospitalization Readmissions

Hospital readmissions for acute exacerbation of COPD in patients with previous hospitalization is extremely common. In an audit of 1400 patients hospitalized for an acute exacerbation of COPD, the authors found that 2/3 of patients had a previous hospitalization [[61]]. A prospective study of patients with previous admission for acute exacerbation of COPD found that 63% of patients were readmitted in 1 year of follow‐up [[33]]. In general, most hospital readmissions are believed to be caused by unmodifiable causes such as patient frailty and progression of chronic disease. However, a meta‐analysis [[104]] on global readmissions concluded that as much as 55% of early readmissions could be due to poor quality and theoretically modifiable care during index hospitalization, such as poor resolution of the main problem, unstable therapy at discharge and inadequate post discharge care [[93]][[104]]. The reasons for hospital readmissions for AECOPD await further clarification from randomized controlled interventional trials specific to COPD readmission.

Quality of Life

The functional and health status [quality of life] of patients with COPD in the months following hospital discharge for an acute exacerbation of COPD is markedly impaired. These patients are more likely to be dependent on others for basic functions, to have impaired physiological reserve and to be limited by disabling symptoms. Half of these patients admit to having only fair to poor quality of life, a number which is five times that in the general population [[5]].

There is little longitudinal data to show whether exacerbations cause a temporary or progressive decline in QOL in the natural history of the disease. Nevertheless, the results of studies which evaluated the QOL showed that patients with COPD who have more frequent exacerbations have worse quality of life [[105]] and poor QOL score is associated with frequent exacerbations and hospitalizations [[6]][[9]][[13]].

In a prospective cohort study of primary care clinics of Dept. of VA medical centers [[50]], Fan et al. identified 5503 patients with chronic lung diseases, from 24,458 patients enrolled in general internal medicine clinics who completed a health inventory. 3,282 patients with self‐reported lung disease, completed the baseline of the Seattle obstructive lung disease questionnaire [SOLDQ], a COPD‐specific QOL questionnaire, and were followed up for 1 year. They found that 4.3% of the patients were hospitalized for COPD‐related illness. The relative risk of any hospitalization for patients in the lowest quartile score was double those with the highest quartile score. Patients with a QOL score in the lowest quartile had a hospitalization score for COPD that was 5 times that in the upper quartile. The authors concluded that lower QOL score [the lower the score, the worse the QOL] is a powerful predictor of hospitalization and an all‐cause mortality in COPD, and suggested that the QOL could be a useful tool for identifying patients who could benefit from preventive measures (Figure 3).

Figure 3. Health related quality of life can predict hospitalization and mortality in patients with COPD. (From Ref. [[50]].)

In another prospective study in the ambulatory care setting [[13]], a cohort of 70 physician‐diagnosed, well‐documented COPD patients, in which exacerbations were carefully defined on diary cards, Seemungal et al. found that the cohort had a median of 3 exacerbations in 1 year, of which 50% were not reported. The patients were grouped as those who had less than 3 [infrequent] and those who had more than 3 [frequent] exacerbations. The total and component scores of the health‐related quality of life using the St. George's questionnaire [SGRQ] were worse in the frequent exacerbators, suggesting that QOL scores may be used to predict frequent exacerbations.

In a prospective cohort study [[6]] in which patients were recruited from a hospital setting, Osman et al. followed up 266 patients who had been hospitalized for exacerbations of COPD for 1 year to determine whether QOL scores could prospectively predict re‐admission for COPD or death within 12 months of an original admission. The authors found that higher (worse) scores on the SGRQ were significantly related to re‐admission for COPD in the next 12 months and concluded that poor (high) scores on the SGRQ, which measures patient distress and coping, are independently associated with re‐admission for COPD and additional use of resources such as nebulizers.

Strategies for Prevention or Reduction of Exacerbations/Hospitalization and Rehospitalization

Conventionally, tertiary health care areas such as emergency room and hospitalization are organized to provide efficient “crisis management” and very little attention is focused on preventing a repeat of the crisis. As a result, patients may leave the hospital less equipped to cope with independent life than before admission. The huge societal and financial cost of hospitalizations has triggered interventional studies into methods of preventing or reducing exacerbations and in particular hospitalization.

The clinical objectives of intervention strategies in COPD are: 1) the reduction of the frequency or severity of exacerbations, or hospitalizations; 2) the amelioration of quality of life deterioration; and 3) the modification of the natural history of the disease. The outcomes of these interventions or ‘altered exposure’ are conventionally studied by the clinical trials. Epidemiologically, the clinical trial design of intervention studies can also be used to determine whether a link between a suspected risk factor and disease is causative rather than simply an association. Hence interventional studies are the preferred method for the study of risk factors of exacerbation and hospitalizations of COPD.

There are many exacerbation‐associated factors identified by retrospective and prospective studies but confirmation of risk factors awaits results from intervention studies. Meanwhile, a review of the existing clinical trials which measure COPD exacerbations as an outcome may provide further insight into putative risk factors for exacerbation and hospitalizations for COPD. Conversely, a better understanding of risk factors would further improve intervention strategies for the management of COPD.

Exacerbations as Outcome in Clinical Trials

To date, there are no disease‐modifying drugs available that can change the natural history of COPD and smoking cessation [[106]] remains the only intervention that has clearly been shown to be effective in slowing the accelerated decline in lung function. Neither anticholinergic drugs, investigated in the lung health study [[106]] nor large scale clinical drug trials of inhaled corticosteroids [[44]][[45]][[46]] could slow the rate of decline in lung function. However, when other outcomes were evaluated these drugs had a significant impact on the frequency of exacerbation, clinical and lung function recovery from exacerbation, and quality of life deterioration [[23]][[24]][[25]][[26]][[107]].

Intervention strategy needs to be specific. Non‐specific interventions such as increased access to primary care [[54]][[108]][[109]], that do not target specific risk factors are unlikely to succeed in reducing hospital admissions. In a secondary analysis of a randomized control trial to evaluate increased access to primary care in common chronic diseases, Smith et al. found that 28 to 38% of COPD patients, depending on the severity, were readmitted in 90 days [[54]]. The authors also found that increased access to primary care and better patient satisfaction did not reduce hospitalization; rather, the factors were associated with increased rates of readmission. This paradoxical effect on readmission emphasized the need to differentiate between increased access to health care services and a real improvement in the quality of care.

Interventions directed at modifiable risk factors such as functional status in pulmonary rehabilitation [[110]], viral infection in influenza vaccination [[55]], and undertreatment in specific pharmacological treatments [[25]][[26]][[29]][[107]] have been found to be variably effective in reducing exacerbations of COPD in randomized trials.

However, as multiple risk factors for exacerbations may exist, and are often interrelated, a single approach is unlikely to be useful. The diverse factors associated with advanced COPD demand a multidisciplinary approach by a professional team [[34]][[58]][[110]], and both pharmaceutical and non‐pharmaceutical approaches play major roles. The strategies for prevention of exacerbations and in particular hospitalizations are therefore complex.

Several specific strategies have been shown to prevent or reduce exacerbations of COPD. Immunization against influenza has been shown to cause a 50% reduction in mortality due to exacerbations. Inhaled corticosteroids have been found to reduce exacerbation frequency by 25% in patients with moderate/severe disease [[23]][[24]][[25]]. The circumstances for the use of antibiotic therapy are now established in patients with increased breathlessness, increased sputum production and/or sputum purulence. In those with respiratory failure, noninvasive ventilation has been shown to reduce intubation rates, shorten lengths of hospitalization, and improve mortality. Early or immediate supported discharge for selected patients has been shown to be effective in the management of patients with COPD [[111]][[112]][[113]].

Rehabilitation

Pulmonary rehabilitation is poorly provided for in most countries: 3% in a UK study [[62]] and 14% in a Spanish study [[32]]. The benefits of pulmonary rehabilitation are well established and include increased exercise capacity, improved functional status, quality of life, and survival [[34]][[103]]. Trials of pulmonary rehabilitation which improved exercise capacity have shown a reduction in duration of hospitalization [[114]] but not a reduction in hospital admissions. Whether pulmonary rehabilitation can reduce readmission for COPD exacerbation remains to be tested in interventional trials.

Pharmacological Agents

Inhaled corticosteroids have become the recommended maintenance treatment for COPD patients with updated GOLD classification [[34]] Stage III [severe] and Stage IV [very severe] and repeated exacerbations based on the results of 4 randomized control trials which showed reduction in the frequency of exacerbations and improvement in health status [[23]][[26]]. However, it is unclear whether this group of drugs would also reduce hospitalizations [[115]][[116]][[117]][[118]].

A recent metaanalysis [[119]] concluded that long‐acting beta2‐agonists and anticholinergics (tiotropium) reduced exacerbation rates by approximately 20% to 25% in patients with moderate to severe COPD. Furthermore, combining a long‐acting beta2‐agonist with an inhaled corticosteroid resulted in even greater reduction in exacerbations. However, the two main concerns persist: 1) in the definition of exacerbations as onset or worsening of symptoms of cough, dyspnoea, or wheeze, making it difficult to exclude the possibility that results may just be due to the pharmacological properties of the bronchodilators, and 2) due to the fact that many patients were simultaneously treated with corticosteroids.

Primary/Home Care

Because of the cost of hospitalization and the poor outcomes of patients hospitalized for COPD, the alternative to hospitalization for patients with acute exacerbation of COPD has been the subject of several recent studies. The specific exacerbation‐related issues addressed were: 1) the optimal length of hospitalization and early hospital discharge with home support [[111]][[112]][[113]][[120]][[121]]; 2) home management of an acute exacerbation [[121]][[122]][[123]][[124]].

Early Discharge with Home Support

Early discharge of a patient hospitalized for acute exacerbation of COPD can result in an increased risk of re‐hospitalization [[104]][[125]]. The optimal length of hospitalization for acute exacerbation of COPD is unclear. In a prospective study by Mushlin Ai et al. [[120]] to determine the medically‐directed length of stay compared to that statistically assigned by the diagnostic related group [DRG] system, the needed length of stay of the patients was longer than the DRG assigned length. The GOLD guidelines recommend that patient should be discharge according to discharge criteria based on limited data and consensus (Table 13) and suggested items for follow‐up assessment 4–6 weeks after discharge from hospital. There is some evidence from randomized trials that follow‐up home visits by a specialized community nurse may permit earlier hospital discharge of COPD patients, without increasing readmission rate [[111]][[112]][[113]][[121]]. However, more information is needed on the extent of home care and the patient selection criteria.

Table 13.  Global Initiative for COPD [GOLD] Consensus for hospital Discharge

Discharge criteria for patients with exacerbations of COPD
• Inhaled β2‐agonist therapy is required no more frequently than every 4 hrs.
• Patient, if previously ambulatory, is able to walk across room.
• Patient is able to eat and sleep without frequent awakening by dyspnea.
• Patient has been clinically stable for 12–24 hrs.
• Arterial blood gases have been stable for 12–24 hrs.
• Patient (or home caregiver) fully understands correct use of medications.
• Follow‐up and home care arrangements have been completed (e.g., visiting nurse, oxygen delivery, meal provisions).
• Patient, family, and physician are confident patient can manage successfully.

Home Management

Home ‘hospitalization’ has been studied as an alternative to conventional hospitalization. There are some data from randomized trials [[121]][[122]][[123]][[124]][[126]] that patients with acute exacerbation of COPD but without acidotic respiratory failure can be effectively managed at home, reducing the need for hospitalization. However, the exact criteria for determining the optimal type of patients for home management have not been clearly defined, and its cost effectiveness in different settings is unknown.

Conclusions and the Future

Acute exacerbations of COPD and resultant hospitalizations are common, occur with increasing severity of the disease, are associated with considerable morbidity and mortality and are responsible for the main cost of health care for the disease.

There is an urgent need for a consensus definition or a choice of definitions and or markers of AECOPD for research purposes and to facilitate inter‐study comparisons. There is little information on the causes and mechanisms of community exacerbations and the relationship between these AECOPD and those that require hospitalizations. Many potential risk factors have been proposed in prospective studies but these need confirmation by randomized interventional studies. Such studies on strategies to reduce severe exacerbations of COPD and hospitalizations allow both the evaluation of the outcomes of the different interventions and the determination of the causative importance of putative risk factors. The optimal strategy for the reduction of hospitalization in COPD remains unclear, though there is some evidence to suggest that a multi‐modality approach consisting of a continuum of self‐management is effective in reducing recurrent hospitalizations. Further long‐term interventional studies are needed.