Reasons for ICU Demand and Long-term Follow-up of a Chronic Obstructive Pulmonary Disease Cohort

Abstract

Acute respiratory failure (ARF) can necessitate mechanical ventilation and intensive care unit (ICU) admission in patients with COPD. We evaluated the reasons COPD patients are admitted to the ICU and assessed long-term outcomes in a retrospective cohort study in a respiratory level-III ICU of a teaching government hospital between November 2007 and April 2012. All COPD patients admitted to ICU for the first time were enrolled and followed for 12 months. Patient characteristics, body mass index (BMI), long-term oxygen therapy (LTOT), non-invasive ventilation (LT-NIV) at home, COPD co-morbidities, reasons for ICU admission, ICU data, length of stay, prescription of new LTOT and LT-NIV, and ICU mortality were recorded. Patient survival after ICU discharge was evaluated by Kaplan–Meier survival analysis. A total of 962 (710 male) patients were included. The mean age was 70 (SD 10). The major reasons for ICU admission were COPD exacerbation (66.7%) and pneumonia (19.7%). ICU and hospital mortality were 11.4%, 12.5% respectively, and 842 patients were followed-up. The new LT-NIV prescription rate was 15.8%. The 6-month 1, 2, 3, and 5-year mortality rates were 24.5%, 33.7%, 46.9%, 58.9% and 72.5%, respectively. Long-term survival was negatively affected by arrhythmia (p < 0.013) and pneumonia (p < 0.025). LT-NIV use (p < 0.016) with LTOT (p < 0.038) increase survival. Pulmonary infection can be a major reason for ICU admission and determining outcome after ICU discharge. Unlike arrhythmia and pneumonia, LT-NIV can improve long-term survival in eligible COPD patients.

Keywords :

• chronic obstructive pulmonary disease
• intensive care unit
• long-term mechanical ventilation
• reason for respiratory failure

Introduction

Chronic obstructive pulmonary disease (COPD) is a leading cause of morbidity and mortality worldwide, and its prevalence is expected to increase to be the third leading cause of death in 2020 (1, 2). Patients with advanced stages of COPD can experience respiratory failure that requires oxygen and mechanical ventilation (non-invasive or invasive), both short term in the hospital/intensive care unit (ICU), and long term at home (3–5).

Acute COPD exacerbation is a major cause of acute respiratory failure (ARF) that leads to ICU admission. Studies researching COPD patient outcome following ICU admission and a long-term ICU stay have largely excluded any other primary reasons, other than acute exacerbation of COPD (6,7,8). There are reports on ICU outcomes of COPD patients with pneumonia (9, 10); however, there is limited data on these COPD patients regarding all causes for ICU demand and long-term stay, and mortality after ICU discharge. In the present study, we aimed to evaluate the reasons for ICU admission and long-term outcome after ICU discharge. Factors affecting mortality in the ICU were not assessed.

Methods

We designed a retrospective, observational, cohort study in a 22-bed, level III ICU of a tertiary teaching hospital for chest diseases, between November 2007 and April 2012. During the study period, a total of eight intensivist specialists worked in the ICU, which they staffed 24 hours a day. This study was approved by the local ethics committee of Sureyyapasa Chest Diseases and Thoracic Surgery Teaching and Research Hospital, Istanbul, Turkey. Ethical approval was in accordance with the Declaration of Helsinki. Due to the retrospective nature of the study design, informed consent was not obtained as half of the patients had died at the beginning of study.

Patients

All consecutive patients with previously diagnosed COPD, who were admitted to our ICU due to ARF and were followed-up for more than 12 months, were included. The COPD diagnosis was established by a physician who evaluated airflow obstruction on spirometry, i.e. forced expiratory volume in 1 second (FEV₁) of 70% predicted or less, and an FEV₁ and forced vital capacity ratio of 70% or less (11). Spirometry test data were not recorded from the patients’ charts. Only those patients who were admitted to the ICU for the first time with ARF were evaluated in case of recurrent admissions to the ICU during the study period.

Data

Demographics and reasons for ICU admission associated with respiratory failure, such as sepsis/septic shock, pneumonia, pulmonary embolism, and pneumothorax, were recorded from the patients’ ICU files (5, 12, 13). Co-morbidities including diabetes mellitus, arrhythmia (i.e. atrial fibrillation), hypertension, congestive heart failure, coronary artery disease (CAD), and depression were recorded (14, 15). History of smoking, use of long-term oxygen therapy (LTOT), and long-term mechanical ventilation (non-invasive (NIV) or invasive (IMV)) via tracheostomy were also recorded. The acute physiology and chronic health evaluation II (APACHE II) score was calculated as part of the patients’ ICU severity index and on admission to the ICU (16). Arterial blood gas (ABG) values at the time of ICU admission and discharge from ICU were recorded from the patients file. Serum C-reactive protein (CRP) levels, the application of NIV and/or IMV in the ICU, the length of ICU and hospital stay (days), and ICU mortality were recorded on admission to, and discharge from the ICU.

Definitions

Hypoxic ARF was defined as the ratio of partial arterial oxygen pressure to inspired fractionated oxygen (PaO₂/FiO₂) < 300 and partial arterial carbon dioxide pressure (PaCO₂) < 45 mmHg. Hypercapnic/hypoxemic ARF was defined as PaCO₂ > 45 mmHg and PaO₂/FiO₂ < 300, and hypercapnic ARF was PaCO₂ > 45 mmHg and PaO₂/FiO₂ > 300 (5, 12). The definition of COPD exacerbation due to an infectious origin was defined by the presence of all three Anthonisen's criteria as follows: worsening of dyspnea, increased volume of pulmonary secretions (endotracheal, sputum), and increased purulence of respiratory secretions (11,17).

Mechanical ventilation

Initially, NIV was applied to all COPD patients with hypercapnic respiratory failure except when absolutely contraindicated (5, 12, 18). Contraindications of NIV were defined as: 1) absolute, respiratory arrest and unable to fit mask, and 2) relative, medically unstable (hypotensive shock, uncontrolled cardiac ischemia or arrhythmia, uncontrolled copious upper gastrointestinal bleeding), agitation, uncooperativeness, inability to protect airway, impaired swallowing, excessive secretions not managed by secretion clearance techniques, multiple (two or more) organ failure, and recent upper airway or upper gastrointestinal surgery (5, 12, 18).

NIV was provided in pressure assist-control mode with ICU mechanical ventilators via a double-tube circuit with a full-face mask. Pressure support (PS) was initially set at 8–10 cmH₂O and gradually increased to a maximum of 30 cmH₂O until the exhaled tidal volume was 5–7 mL/kg and guided by patient tolerance. Positive end-expiratory pressure (PEEP) was set at 5 cmH₂O and raised or lowered to treat hypoxemia or enhance patient comfort, respectively. FiO₂ was adjusted to maintain oxygen saturation (SaO₂) at 90%. NIV was applied intermittently for periods of 1 to 4 hours, and initial ABG samples were obtained at the end of the first hour. The duration of each session was determined by ABG value improvement, consciousness level, and patient compliance. The definition of NIV failure in hypercapnic patients was no pH improvement, no change or a rise in breathing frequency after 1–2 hours, and lack of cooperation. For hypoxic COPD patients, failure was considered as no, or a minimal, rise in PaO₂/FiO₂ after 1–2 h (< 200) (5).

IMV was applied in the presence of absolute or relative contraindications for NIV, as mentioned above. IMV was applied by assist control ventilation (A/C) in pressure control mode. Inspiratory pressure was set for a target tidal volume of 5–7 mL/kg ideal body weight (airway plateau pressure < 30 cmH₂O) under a sedation protocol. The Richmond agitation sedation scale (RASS) was used for infusion and assessment of the daily need for sedation (19). When patients met the previously described criteria for weaning, the pressure support ventilation (PSV) mode was used and gradually decreased (1–2 cm H₂O every 1–2 hours) (20). When the PS reached 8–10 cm H₂O and PEEP was 0 or < 5 cm H₂O the patient progressed to spontaneous breathing trials (SBT) using a T-piece with oxygen support; the T-piece trial duration was 30 minutes, after which time they were extubated. NIV was then applied in cases of moderate respiratory distress if there was no contraindication (20).

Bronchodilator and anti-inflammatory treatment in the ICU (11, 21, 22)

A short-acting ß2 agonist (salbutamol, 100 mcg per puff) and ipratropium bromide (100 mcg/20 mcg per puff) were given every 2–4 hours (4 to 10 puffs) via a metered dose inhaler chamber (Aerovent, Altech®, Altera Firm, Izmir-Turkey) when the patients were under NIV or IMV. A nebular form of salbutamol (2.5 mg/2.5 mL per nebule) was given every 15 minutes to 4 hours, or ipratropium bromide/salbutamol (0.5 mg/3.01 mg/2.5 mL per nebule) was given every 2 to 4 hours for patients breathing without mechanical ventilation support.

Long-acting ß2 agonists were not used in COPD patients with ARF in the ICU. Intravenous methylprednisolone (40–60 mg) was given one to two times daily as an anti-inflammatory for those patients unable to take oral medications or with impaired gastrointestinal absorption. The steroid dose was tapered gradually and discontinued over 7 to 10 days. Methylxanthines (theophylline and aminophylline) were not given. All patients received oxygen for COPD, as well as medication to treat the underlying cause of ICU admission or ARF and co-morbidities, such as antibiotics and anti-arrhythmia or anticoagulant therapies (11).

Follow-up

Long-term NIV was prescribed according to the national guidelines (based on international guidelines) for COPD patients after ICU discharge with day time hypercapnia (e.g. PaCO₂ greater than 50 mmHg) and oxygen desaturation during sleep (e.g. SpO₂ ≤ 88% for ≥5 minutes of ≥ 2 hours of nocturnal sleep oximetry), despite the use of supplemental oxygen at ≥2 L/min. It was also prescribed to those with an acute exacerbation that necessitated the use of continuous NIV in hospital more than two times per year (23). A spontaneous bi-level positive airway pressure NIV device was prescribed if the need for inspiratory pressure was greater than 20 cmH₂O, or in the presence of ineffective respiratory effort/apnea in the ICU. LTOT was prescribed when PaO₂ was less than 55 mmHg or SaO₂ was 88% or less in room air at the time of discharge from ICU, and when hypoxemia had occurred for 1 to 2 months prior, in a clinically stable period (11). The prescription of LTOT and NIV devices after ICU discharge, and long-term survival were recorded from the patient's files. The time of death was determined from outpatient files and online in May 2013.

Statistical analysis

A descriptive analysis was used to investigate patient demographics and ICU data. Groups were compared with Mann Whitney U-tests for non-parametric continuous variables or Student's t-tests for parametric continuous variables. Chi-square tests were employed for dichotomous variables. The median with interquartile range (IQR) was employed for non-parametric continuous variables, and mean ± standard deviation (SD) was used for parametric continuous variables. Count and percentage were used when applicable.

To predict the long-term mortality of COPD patients after ICU discharge, Kaplan-Meier survival analyses were performed. Patients’ co-morbidities, reasons for ICU admission, CRP levels, APACHE II score on admission, PaCO₂ and PaO₂/FiO₂ at ICU discharge, HCO₃, prescription of LT-NIV at home, and LTOT after ICU discharge were included as variables. Long-term mortality risk factors were analyzed with the Cox regression model. We included the variables that were of statistical significance following univariate analyses of survival and non-survival in patients following ICU discharge. A p value < 0.05 was accepted as statistically significant.

Results

Patients

During the study period, 3371 patients were admitted to the ICU, and 1264 admissions (including re-admission) were due to ARF in COPD patients. Only those who were admitted to the ICU were evaluated and, as such; 962 patients were ultimately included in the study. Inclusion criteria are summarized in Figure 1.

Figure 1.   Inclusion and follow-up flow chart.

Patient demographics (age, gender, body mass index (BMI), history of cigarette smoking), co-morbidities (diabetes mellitus, hypertension, arrhythmias, coronary artery disease (CAD), depression) and baseline data of utilization of LTOT and LT-NIV or IMV before ICU admission are shown in Table 1. The majority of COPD patients were male, elderly (age > 65 years), and overweight.

Table 1.  Characteristics of chronic obstructive pulmonary diseases patients with acute respiratory failure in the intensive care unit

Variables	Value
Patients number, n	962
Age, year, mean ± SD	70 ± 10
Gender, Female/Male	252/710
Body mass index, kg/m^2, mean ± SD	26 ± 6
Cigarette smoking, n (%)	752 (78.2)
Biomass exposure, n (%)	210 (21.8)
Co-morbid diseases, n (%)
Diabetes mellitus	167 (17.4)
Hypertension	317 (33.1)
Arrhythmia	113 (11.8)
Coronary arterial disease	109 (11.4)
Depression	29 (3.0)
Long-term home device before ICU, n ( %)
Non-invasive mechanical ventilation	120 (12.5)
Oxygen therapy	352 (36.6)

SD: standard deviation, ICU: intensive care unit

ICU data

ICU data including reasons for ICU demand (acute COPD exacerbation, pneumonia, embolus, pneumothorax), APACHE II score on ICU admission, CRP and ABG values on ICU admission and discharge, length of ICU stay, and ICU mortality are shown in Table 2.

Table 2.  Intensive care unit data of chronic obstructive pulmonary diseases with acute respiratory failure

Reasons of ICU admission, n(%)
Acute exacerbation of COPD	640 (66.7)
Pneumonia	189 (19.7)
Pulmonary embolism	43 (4.5)
Pneumothorax	13 (1.4)
APACHE II score on admission to the ICU, mean ± SD	20 ± 7
C-reactive protein, mg/L, at the time of admission to the ICU, median	32.4 (14.0–80.0)
C-reactive protein, mg/L, at the time of discharge from ICU, median	21.9 (11.3–54.2)
Arterial Blood Gases, at the time of admission to the ICU
pH, mean ± SD	7.30 ± 0.31
PaCO2, mmHg, mean ± SD	72.1 ± 21.3
PaO2/FiO2, median (IQR)	181 (130–254)
SaO2 %, median (IQR)	92.8 (84.0–96.8)
HCO3 mmol, mean ± SD	34.1 ± 12.5
Arterial Blood Gases, at the time of discharge from ICU
pH, mean ± SD	7.43 ± 0.09
PaCO2, mmHg, mean ± SD	54.4 ± 13.8
PaO2/FiO2, median (IQR)	233 (190–297)
SaO2 %, mean ± SD	95.5 ± .30
HCO3 mmol, mean ± SD	34.7 ± 7.2
Tracheostomy procedure at the ICU	42 (4.4)
Length of stay in the ICU day median (IQR)	6 (2–10)
Patients with only medical treatment, day median (IQR)	2 (1–4)
Patients with only NIV application, day median (IQR)	6 (8–9)
Patients with only IMV application, day median (IQR)	6 (2–13)
Patients with IMV after NIV failure, day median (IQR)	10 (6–15)
Mortality in the ICU n (%)	110 (11.4)
Mortality in the hospital n (%)	120 (12.5)
Length of hospital stay, day median (IQR)	12 (8–18)

ICU: intensive care unit, COPD: chronic obstructive lung diseases, APACHE II: acute physiological and chronic health evaluation II, SD: standard deviation, IQR: inter quartile range, PaCO₂: partial arterial carbon dioxide pressure, PaO₂/FiO₂: partial arterial oxygen pressure over fractionated inspired oxygen pressure, SaO₂: arterial oxygen saturation, HCO₃: Bicarbonate. LTOT: Long-term oxygen therapy.

Nearly all patients were hypercapnic and hypoxemic, and one-fourth of the patients were severely acidotic (pH < 7.25). The majority of COPD patients with ARF were admitted to the ICU due to an infection, including pneumonia (86.4%). The pathogens responsible for the infections were not recorded. Airway infection was observed in 66.7% of patients without lung parenchymal infiltration. Patients with pneumonia (n = 189, 19.7%) required initial IMV (n = 17) and NIV (n = 167). Among pneumonia patients who initially received NIV, the intubation rate i.e., NIV failure, was 32.3% (n = 54). The overall mortality rate of patients with pneumonia was 21.7% (n = 41) in the ICU; but the mortality rate of pneumonia patients with IMV, NIV failure and those who only used NIV were 7/17, 18/54 and 13/113, respectively. Three patients with pneumonia who did not consent to intubation were only treated medicinally and died in the ICU. One patient with pneumonia died in the ward after ICU discharge.

Figure 2 shows the treatment approach and mortality of COPD patients with ARF. Although 86% of COPD patients were treated initially with NIV, initial IMV was applied in only 8.6% of cases. Overall, intubation was performed in 24.3% of patients, including those with NIV failure. The mortality rate of patients with NIV followed by IMV (NIV failure), those with NIV success, and those with initial IMV use, were 15%, 8%, and 28%, respectively, and were significantly different from each other (p < 0.001). ICU and hospital mortality rate were 11.4% and 12.5%, respectively.

Figure 2.   Treatment approaches and long-term survival of COPD patients in the ICU.

Long-term Survival

After ICU discharge, 842 patients were followed-up for a minimum of 12 months and a maximum of 64 months. The median and IQR of survival was 15 months (3 to 29 months). During follow-up, the 3- and 6-month mortality rates were 17.1% and 24.5%, respectively (Figure 3). The numbers of and mortality rates of patients followed for 1, 2, 3, 4, and 5 years were n = 842 (33.7%), n = 612 (46.9%), n = 393 (58.9%), n = 214 (64.4%), and n = 80 (72.5%), respectively (Figure 3).

Figure 3.   COPD patient survival after ICU discharge shown in months as determined by Kaplan–Meier analysis.

Table 3 demonstrates the utilization of long-term oxygen and noninvasive mechanical ventilation therapy in COPD patients with acute respiratory failure after ICU discharged.

Table 3.  The utilization of long-term oxygen and noninvasive mechanical ventilation therapy in COPD patients with acute respiratory failure after ICU discharge

Discharged patients, n	842
LTOT, total n (%)	407 (48.3)
Previously LTOT device, n (%)	300 (35.6)
Newly prescribed LTOT device, n (%)	115 (13.6)
Long-term NIV at home, n (%)	241 (28.6)
Previous long-term NIV device, n (%)	108 (12.8)
Newly prescribed long-term NIV device, n (%)	133 (15.8)
Long-term tracheostomized-IMV at home, n (%)	41 (4.9)

LTOT: long-term oxygen therapy, NIV: noninvasive mechanical ventilation, IMV: Invasive mechanical ventilation.

Table 4 shows a a univariate analyzes of those patients who survived (n = 363) and those who did not survive (n = 479) in the follow-up period. Those patients who did not survive in the long-term were older, had a lower BMI, higher APACHE II score on admission to the ICU, and a higher rate of IMV use after NIV failure than survivors (Table 4). Long-term survivors had a statistically significant higher rate of NIV and shorter length of ICU and hospital stay than non-survivors.

Table 4.  The univariate analyzes of survival and non-survival COPD patients’ characteristics and ICU data after hospital discharge

	Survival, n = 363	Non-survival, n = 479	p values
Female/Male	104/259	120/359	0.24
Age, year, n (%)
≥ 80	41 (11.3)	95 (19.8)	0.001
79–65	173 (47.7)	282 (58.9)
≤64	149 (41.0)	102 (21.3)
Co-morbidities, n (%)
Hypertension	129 (35.5)	154 (32.2)	0.31
Ischemic heart diseases	37 (10.2)	57 (11.9)	0.43
Diabetes mellitus	60 (16.5)	79 (16.5)	0.99
Arrhythmia	33 (9.1)	59 (12.3)	0.14
Depression	9 (2.3)	15 (3.1)	0.57
Body mass index, kg/m^2	n=322	n=420
≤20, n (%)	38 (11.8)	76 (18.1)	0.001
21–25, n (%)	114 (35.4)	192 (45.7)
26–30, n (%)	93 (28.9)	106 (25.2)
31–39, n (%)	60 (18.6)	32 (7.6)
≥40, n (%)	17 (5.3)	14 (3.3)
Long-term oxygen therapy, n (%)
Previously ICU admission	134 (36.9)	166 (34.7)	0.50
After ICU discharge	56 (15.4)	51 (10.6)	0.039
Long-term home NIV, n (%)
Previously ICU admission	48 (13.2)	60 (12.5)	0.77
After ICU discharge	72 (19.2)	61 (12.7)	0.005
Exposure to smoke, n (%)
Cigarette	282 (77.7)	378 (78.9)	0.67
Biomass	81 (22.3)	101 (21.1)
ICU data
APACHE II score on admission to the ICU
≥ 30, n (%)	11 (3.0)	39 (8.1)	0.001
29–20, n (%)	120 (33.1)	188 (39.2)
19–16, n (%)	111(30.6)	138 (28.8)
≤ 15, n (%)	121 (33.3)	114 (23.8)
Respiratory failure, n (%)	361 (99.4)	477 (99.8)	0.41
COPD exacerbations (infections)	231 (63.9)	315 (65.9)	0.55
Pneumonia	53 (14.6)	94 (19.7)	0.055
Pulmonary embolism	16 (4.4)	23 (4.8)	0.78
Pneumothorax	5 (1.4)	6 (1.3)	0.88
Mechanical ventilation in the ICU, n (%)
Invasive ventilation	19 (5.2)	40 (8.4)	0.08
Invasive ventilation after noninvasive failure	58 (16.0)	110 (23.0)	0.015
Noninvasive ventilation	268 (73.8)	307 (64.1)	0.003
Medical treatment only	18 (5.0)	22 (4.6)	0.81
Tracheostomies	11 (3.0)	30 (6.3)	0.031
Arterial blood gases, discharge from ICU, median (IQR)
pH	7.44 (7.41–7.48)	7.43 (7.39–7.47)	0.001
PaCO2 mmHg	54.0 (47.0–59.6)	53.9 (46.9–60.7)	0.94
PaO2/FiO2	231(192–288)	235 (188–297)	0.82
HCO3 mmol	35.9 (31.9–39.1)	34.4(30.2–38.2)	0.15
C-reactive protein mg/dL, discharge from ICU, median (IQR)	18.5 (9.5–46.5)	21.5 (11.2–47.9)	0.45
Length of ICU stay, day, median (IQR)	6 (4–9)	7 (4–11)	0.007
Length of hospital stay, day median (IQR)	12 (8–16)	13 (9–20)	0.001

ICU: intensive care unit, COPD: chronic obstructive lung diseases, APACHE II: acute physiological and chronic health evaluation II, SD: standard deviation, IQR: interquartile range, PaCO₂: partial arterial carbon dioxide pressure, PaO₂/FiO₂: partial arterial oxygen pressure over fractionated inspired oxygen pressure, HCO₃: Bicarbonate.

Gender, co-morbidities, reasons for ICU admission (pneumonia, pulmonary embolus, pneumothorax, COPD exacerbation), and utilization of LTOT and LT-NIV after ICU discharge, were analyzed with Kaplan–Meier analyses to identify factors affecting long-term survival. We observed statistically significant shorter survival in COPD patients with arrhythmia or pneumonia after ICU discharge (p < 0.013 and p < 0.025, respectively) (Figure 4a, Figure 4b).

Figure 4a.  (a) Kaplan–Meier survival curves of COPD patients with arrhythmia after ICU discharge. Patients with arrhythmia died earlier than patients without arrhythmia after ICU discharge (p < 0.013). (b) Kaplan–Meier survival curves of COPD patients with pneumonia as a reason for ARF after ICU discharge. Patients with pneumonia died earlier than those without (p < 0.025).

LT-NIV and LTOT prescription after ICU discharge were associated with significantly increased survival of COPD patients (p < 0.016 and p < 0.038, respectively) (Figure 5a, b). Figure 5c shows the survival of COPD patients with LT-NIV at home (previously, plus newly prescribed) after ICU discharge. Patients with LT- NIV after ICU discharge appeared to live longer than others, but this result did not quite reach significance (p = 0.052).

Figure 5.  (a) Kaplan–Meier survival curve of COPD patients with new LTOT use after ICU discharge. Those with LTOT lived longer than patients without (p < 0.038). (b) Kaplan–Meier survival curve of COPD patients with new LT-NIV use at home after ICU discharge. Those who used LT-NIV after ICU discharge lived longer than patients who did not (p < 0.016).

Figure 5.  (c) Kaplan–Meier survival curve of COPD patients with LT-NIV at home (previously plus newly prescribed) after ICU discharge. Patients with LT- NIV after ICU discharge lived longer than others who did not use LT-NIV (p = 0.052).

The Cox regression model was used to analyze risk factors for mortality. The variables included were chosen from those described in Table 4 that were statistically significant, as well as some variables that were not significant but that were suspected to be potentially significant. Thus, the variables analyzed were age, body mass index, gender, smoke history, length of ICU and hospital stay, co-morbidities, pneumonia, type of mechanical ventilation, history of previous and/or new long-term oxygen and NIV at home. The Cox-regression model showed that age above 65 years, ICU admission due to pneumonia, and each additional day in the ICU increased the risk ratio for death.On the other hand, NIV use during the ICU stay and the prescription of LTOT after ICU discharge decreased the risk of mortality during the long-term follow-up (Table 5).

Table 5.  Cox regression analysis of long term survival of COPD patients after ICU discharge

	Risk ratio for death	95% CI lower-upper	p values
Age ≥ 80 years	2.32	1.75–3.07	0.001
Age 65–79 years	1.94	1.54–2.45	0.001
ICU admission due to pneumonia	1.27	1.01–1.59	0.042
Length of hospital stay	1.01	1.00–1.02	0.010
NIV application in the ICU	0.75	0.62–0.91	0.004
LTOT prescription after ICU discharge	0.70	0.52–0.93	0.016

COPD: chronic obstructive pulmonary disease, ICU: intensive care unit, CI: confidence interval, NIV: non-invasive mechanical ventilation, LTOT: Long-term oxygen therapy.

Discussion

This study analyzed the ICU data and long-term outcomes in a large number of COPD patients with ARF. The main reason for ICU admission in these COPD patients was infectious in origin. After ICU discharge, COPD patients with newly prescribed LT-NIV and LTOT lived significantly longer than others. The primary risk factors for long-term mortality after ICU discharge in this group were a longer hospital stay, older age, and a lower BMI. LT-NIV after ICU discharge was associated with decreased long-term mortality.

ICU data

Previous studies investigating respiratory failure in COPD ICU patients mainly looked at acute COPD exacerbation (7), (24). In the present study, two-thirds of patients were admitted to the ICU due to acute COPD exacerbation. Beside COPD exacerbation, pneumonia is an important reason for ICU admission and IMV (25). Rello and coworkers demonstrated the ICU mortality was 39% for COPD patients initially intubated, and 50% for those with failed NIV (25). In the present study, the pneumonia rate was 19.7%, however, the mortality rate was nearly doubled when these patients were intubated initially for IMV, or after NIV failure (37.6%).

Among those cases who received IMV, a 41.2% mortality rate was seen in those who were initially intubated, and 33.3% mortality was observed in those who were intubated after NIV failure. The mortality rate following NIV failure observed here was lower than was reported in Rello's study. Pulmonary embolism is known to be a reason for acute COPD exacerbation. Our data showed that nearly one-fourth of the COPD patients with ARF were severely hypoxemic and had a higher rate of clinical suspicion of pulmonary embolism; and nearly all were intubated for IMV.

NIV is the gold standard for treating hypercapnic ARF in COPD patients (5, 18). The success rate of NIV in this group is between 54% and 85% (7, 26, 27). Confalonieri and colleagues generated a prediction chart that estimates a 70% NIV failure rate if: the pH on admission to ICU is < 7.25, a Glasgow Coma Scale <11, and an APACHE II score >29 (28). In the present study, 86.7% of the COPD patients with ARF had NIV initially applied, and 24.3% of these were intubated. Overall, the NIV failure rate, including those who were intubated and those who were deceased patients during NIV, was 30.8%, which is similar to previously published results. Three fourths of NIV failure patients in our study had blood pH values <7.28, PaCO₂ >76 mmHg, and APACHE II scores >23. These values are in accordance with the chart predicting NIV failure in Confalonieri's study (28).

The mean length of ICU stay was longer for COPD patients with ARF requiring IMV (29). Although we observed similar lengths of ICU stay in both the NIV and IMV groups, a two-fold longer ICU stay was noted for patients who were intubated for IMV after NIV failure. The overall ICU mortality rate was 12.2%, and this was similar to other studies with the exception of Alaitian and coworkers, who recently reported only 6% ICU mortality (7), (29–33).

Long-term follow-up

LTOT is known to be the only treatment proven to increase COPD patient survival (31). However, studies have shown that 60% of COPD patients with LTOT die within 5 years (26, 27), similar to our results. In the ­present study, LTOT was recommended after ICU discharge in 13.6% of cases, according to international guidelines and during follow-up these patients appeared to live longer. LT-NIV has been demonstrated in several studies to improve COPD patient survival. Other randomized studies have reported positive long-term effects (months) of ventilation on CO₂ retention and better quality of life in stable hypercapnic (PaCO₂ >50 mmHg) patients on LTOT and using LT-NIV (34, 35).

These therapies also effectively reduce hospital admissions and minimize costs (36). Casanova and coworkers reported a decreased number of hospital re-admissions, but no positive effect on survival (37). More recently, a randomized control study described a promising effect of prolonged NIV on COPD patient survival of at least 12 months during a 5-year follow-up (38). In the present study, it is important to note that COPD patients with newly prescribed LT-NIV after ICU discharge lived longer than other COPD patients, regardless of whether or not they had previously used NIV.

COPD is a systemic inflammatory disease, and the presence of co-morbidities affects quality of life and long-term survival. Previous reports that evaluated long-term mortality risk factors in COPD patients after ICU discharge failed to identify any risk factors (8, 30). In these studies, the patient numbers were 57 and 74, which are insufficient sample sizes to identify risk factors by logistic regression analyses. In the present study, according to Cox regression model analysis, the long-term mortality risk factors of 842 COPD patients were evaluated over a period of at least 12 months, and 393 patients were included in a 3-year follow-up.

Older age, especially above 80 years, and a longer hospital stay, were identified as risk factors for increased long-term mortality. On the other hand, the prescription of LTOT after ICU discharge and NIV application during the ICU stay were associated with a decreased mortality risk in our study. Kaplan–Meier survival curves revealed that, unlike LT-NIV prescription after ICU discharge, patients with pneumonia or arrhythmia had a shorter survival period after ICU discharge.

There are some limitations to our study. First, it was retrospective and conducted in a single-center. Nonetheless, this large, specific patient group was followed by experienced ICU pulmonologist-intensivists and could provide important results to guide future studies. Secondly, spirometry test scores were not recorded from the patients’ files. Third, no cultures were taken to identify specific infectious pathogens. Finally, our results are not generalizable to all patients. The strengths of this study are the large sample size of a particular group of patients, and the length of the follow-up time.

In conclusion, the primary reasons COPD patients with ARF were admitted to the ICU were acute COPD exacerbation and pneumonia. Initial respiratory support mainly consisted of NIV, but IMV was applied after NIV failure, and a small number of patients were initially ventilated with IMV. Half of the patients survived 2 years after ICU discharge. COPD patients with pneumonia or an arrhythmia had shorter survival times after ICU discharge and, as such, these patients should be closely monitored. The prescription of LT-NIV and LTOT at home prolonged the survival in COPD patients with ARF Thus we suggest prescribing LTOT and LT-NIV to COPD patients with ARF (both hypoxemic and hypercapnic) when discharged from ICU. These findings may be helpful for designing future studies and for a COPD treatment approach.

Declaration of Interest Statement

None of the authors have any conflict of interest, and all contributed to the preparation of the manuscript. The authors alone are responsible for the content and writing of the paper.