Determinants of the Length of Mechanical Ventilation in Patients with COPD in the Intensive Care UnitGursel G.
Department of Pulmonary Diseases, Gazi University School of Medicine, Ankara, Turkey Corresponding Author
Background: About 10% of the patients with chronic obstructive pulmonary disease (COPD) are at high risk for prolonged mechanical ventilation (MV >21 days), and mortality ranges from 55 to 78% in these patients. Objective: To determine the potential risk factors for MV over periods of 1, 2 and 3 weeks in patients with COPD. Patients and Method: The characteristics of patients during the stable period of their disease, on admission to the intensive care unit (ICU) and during the ICU stay were recorded prospectively and analyzed retrospectively for this study. t test, XXX2 test and logistic regression analysis were used for statistical analysis. Results: 86 patients with COPD requiring MV were included in the study. 73, 33, and 13% of the patients required MV longer than 1, 2 and 3 weeks, respectively. There were no significant relationships between the duration of MV and bronchiectasis or the presence of community-acquired pneumonia on admission, baseline pulmonary function test results or blood gas parameters on admission. Development of ventilator-associated pneumonia (VAP; odds ratio, OR: 6; 95% confidence interval, CI: 2–23, p = 0.011) and sepsis (OR: 10; 95% CI: 2–54, p = 0.007) were independent predictors for MV >7 days. VAP was still a risk factor for MV >15 days with an OR of 14 (95% CI: 3–66, p = 0.001). On the other hand MV >21 days was primarily determined by increasing age (OR: 1.2; 95% CI: 1–1.3, p = 0.042), severity of the disease on admission measured by APACHE II score (OR: 1.4; 95% CI: 1–1.7, p = 0.002) and albumin levels (OR: 0.10, 95% CI: 0.01–0.54, p = 0.007). Conclusion: Advanced age, severity of disease on admission and development of VAP during ICU stay are the main determinants of MV duration in patients with COPD.
Copyright © 2005 S. Karger AG, Basel
Chronic obstructive pulmonary disease (COPD) is a major public health problem. It is the fourth leading cause of chronic morbidity and mortality in the United States and is projected to rank five in 2020 as a worldwide burden of disease [1, 2]. For these patients, admission to an intensive care unit (ICU) is common, 26–74% of them receive mechanical ventilatory support, and average hospital and ICU stays are long and expensive [3,4,5]. In studies restricted to patients with COPD and acute respiratory failure, about 10% are at high risk for mechanical ventilation (MV) >21 days, and failure in weaning ranged from 55 to 78%. Weaning failure and prolonged MV (PMV) have been reported as independent predictors of poor outcome in these patients [6,7,8,9]. The major mechanisms of the need for PMV in patients with COPD are the association between abnormal lung mechanics, in particular auto-positive end-expiratory pressure (PEEP), lung resistances, and reduced pressure-generating capacity of the inspiratory muscles because of pulmonary hyperinflation. Many studies have been performed to assess the value of physiological factors as a weaning predictor . Studies of predictors of weaning among patients with COPD have focused on identifying which of those patients already receiving MV are ready to be weaned. However, to our knowledge only in a few studies have other clinical and ICU characteristics been used to predict the length of MV for COPD patients. Early prediction of the length of MV may assist in the early transfer of patients to intermediate care units from ICU and can help to free ICU beds for more acutely ill patients and decrease the cost of care in the ICU.
The aim of this study was to assess if the length of MV in patients with COPD can be predicted from clinical parameters at the time of their admission and during their ICU stay.
Patients and Methods
The study was conducted at a respiratory intensive care unit of a 1,300-bed university hospital during a 2-year period (from January 2000 to December 2001).
Patients with a history and clinical findings of COPD requiring MV subsequently admitted to the respiratory intensive care unit were eligible for participation in this study. The diagnosis of COPD was determined by pulmonary function tests (PFT) when available in 57 of 86 patients (66%) . In the absence of documented airflow obstruction, we used clinical criteria, clinical history with physical findings, evidence of COPD on chest X-ray and patterns of persistent airflow limitation on ventilator flow-volume curves.
Data were collected in all patients requiring MV, and the following information was recorded: baseline characteristics [age, gender, smoking history, comorbidity, association with bronchiectasis, PFT results, classification of severity of airflow limitation, long-term oxygen therapy (LTOT) and home MV therapy (noninvasive), pulmonary artery pressure measured echocardiographically during the stable period of the disease in 20 patients, and the number of prior hospital/ICU admissions], admission characteristics [APACHE II scores, PaO2/FiO2, blood gases, community-acquired pneumonia (CAP) noninvasive MV (NIMV) trial before intubation, systemic steroid use during exacerbations, indications of MV and where the patients were admitted from], and ICU characteristics [length of MV, ICU and hospital stays, requirement of tracheostomy, development of ventilator-associated pneumonia, (VAP) and sepsis, responsible pathogens, day and number of VAPs].
Disease severity was classified according to the criteria of the Global Initiative for Chronic Obstructive Pulmonary Disease (GOLD) . Bronchiectasis was diagnosed with characteristic clinical and radiological findings. Fourteen patients with bronchiectasis and 20 patients without bronchiectasis had computed tomography of the thorax, and for radiological diagnosis we used chest X-rays in other patients. Comorbid conditions recorded were congestive heart failure, hypertension, coronary artery disease, chronic renal failure and diabetes mellitus. Systemic steroid use was defined as intravenous or oral steroid treatment for at least 3 days (30 mg/day).
NIMV was applied first if patients had no intubation criteria. NIMV was applied intermittently, for daily periods of at least 4 h. Pressure support with PEEP was the mode of ventilation. The usual setting was 10–15 cm H2O of pressure support and 0–5 cm H2O of PEEP. Criteria used were similar to the GOLD guidelines for intubation. We classified indications for invasive MV from prospectively recorded data according to the three GOLD criteria: (1) arrest, (2) life-threatening hypoxemia (PaO2 <40 mm Hg or PaO2/FiO2 <200 mm Hg) and (3) severe acidosis (pH <7.25) and hypercapnia (PaCO2 >60 mm Hg) .
Taema ventilators (Horus and Cesar) were used for NIMV and MV, and Moritz II biLEVEL ventilators were used for NIMV in the ICU and pulmonary ward. Some patients were intubated without NIMV in the emergency department.
Protocols for MV and Weaning. All patients were ventilated on assisted control mode at the beginning (tidal volume: 6–8 ml/kg, flow: 30–60 l/min, PEEP: 5 cm H2O, FiO2 adjusted according to oxygen saturation >90%), and when they were able to tolerate this mode they were switched to the pressure support mode. Then the level of pressure support was gradually decreased hourly and after a 4- to 6-hour tolerance of t-tube trial patients were extubated.
VAP was diagnosed based on quantitative endotracheal aspirate cultures yielding >10–5 colony-forming units. For the clinical diagnosis of VAP, the following criteria were used: a new radiographic infiltrate persistent for 48 h or more plus a body temperature >38.5 or <36°C, a leukocyte count of >10 × 109/l or <3 × 109/l, purulent sputum or change in sputum type, and isolation of pathogenic bacteria from an endotracheal aspirate . Both positive clinical and microbiologic diagnoses were necessary for the diagnosis of VAP. First episodes were considered in the risk factor analysis for the duration of MV. For the diagnosis of VAP, the clinical pulmonary infection score was calculated in 17 patients . For the definition of sepsis, 1992 consensus criteria were applied .
Data were reported as means ± SD. The XXX2 test was used to compare percentages. Student’s t test was used for continuous variables. Nonparametric analysis using the Mann-Whitney U test was used for data with non-normal distributions. Logistic regression analysis was used to evaluate the impact of potential risk factors on the duration of mechanical ventilation, controlling for the remaining variables. Odds ratios (OR) and 95% confidence intervals (CI) were calculated in accordance with standard methods. Pearson’s correlation coefficient was used to assess the correlation between the duration of MV and the development of VAP.
MV lasting >1, 2 and 3 weeks were used as the dependent variables in separate analyses. The independent variables used were those found to have ≤0.10 in univariate analysis. Independent variables included in the logistic regression for MV >7 days were age, VAP, sepsis and albumin; for MV >15 days APACHE II, albumin, VAP and sepsis, and for MV >21 days age, APACHE II, albumin, comorbidity, VAP and sepsis. Because of inadequate patient numbers PaO2/FiO2 were excluded from the analysis for MV >7days.
Eighty-six patients with COPD were included in the study. Patients with lung cancer were excluded. 65% (56) of the 86 patients were smokers. 44% (38) of the patients were transferred from the emergency department, 48% (41) from the pulmonary ward, 3% (3) from other ICUs and 5% (4) from other wards. The clinical characteristics of the patients are listed in tables 1, 2, 3. In 9 (10%) patients, tracheostomy had been performed 3 or 4 weeks after the intubation. Admission diagnoses of the patients were acute attack of COPD (55%), CAP (37%), association with obstructive sleep apnea syndrome, severe respiratory acidosis (4%), acute pancreatitis (2%) and pulmonary embolism (2%).
Table 1. Characteristics of the patients and their diseases
Table 2. Patient characteristics on admission
Table 3. ICU data of the patients
The mean duration of MV was 19 ± 24 days, and 73% (63), 33% (28) and 13% (11) of the patients required MV for more than 1, 2 or 3 weeks, respectively. One patient was discharged with tracheostomy and 1 with home NIMV.
There were no significant differences among the patients who were mechanically ventilated over periods shorter or longer than 1, 2 or 3 weeks in PFT parameters, blood gas results, use of LTOT, the diagnosis of bronchiectasis, CAP, use of steroid therapy and prior NIMV (p > 0.05). Albumin levels were significantly lower in patients mechanically ventilated for more than 1 and 3 weeks. PaO2/FiO2 levels were significantly lower in patients ventilated for >1 week compared to those ventilated for <1 week. The number of patients who died or were extubated in each study period are given in table 4.
Table 4. Number of patients deceased or extubated during each period
The length of MV and significantly related parameters are given in table 5. Among these parameters, only VAP and sepsis were independent predictors for the duration of MV >7 days. VAP was also a risk factor for MV >15 days. Increasing age, APACHE II score and decreasing albumin levels were independent predictors for MV >21 days in logistic regression analysis (table 6). Correlation analysis supported these findings; MV >7 days (r = 0.44, p = 0.0001), MV >15 days (r = 42, p = 0.000) and MV >21 days (r = 32, p = 0.002) were significantly but weakly correlated with the development of VAP. 56% of the patients developed 1 ± 1.5 VAP episodes 10 ± 8 days (mean) after intubation (median: 7 day). Responsible pathogens were Pseudomonas aeruginosa (n = 18), Acinetobacter baumanni (n = 12), methicillin-resistant Staphylococcusaureus (MRSA; n = 10), Klebsiella (n = 2), StenotrophomonasDELETEmaltophilia (n = 1), P. aeruginosa and Klebsiella (n = 2), A. baumanni and MRSA(n = 2), and P. aeruginosa and A. baumanni (n = 1).
Table 5. Differences in the clinical characteristics of the patients with different durations of MV
Table 6. Risk factors for the MV > 1, 2 and 3 weeks in patients with COPD
MV >7, 15, and 21 days was significantly associated with increased mortality compared with MV for shorter periods (table 5).
The aim of this study was to evaluate if any clinical characteristic may predict the length of MV in patients with COPD. The most important factors determining MV for >1 week were the development of VAP and sepsis in this study. On the other hand, factors predicting MV for >3 weeks were advanced age, and APACHE II scores and albumin levels on admission. Association with bronchiectasis and presence of CAP on admission were not related to the duration of MV. Lastly, PMV was significantly associated with increased mortality in patients with COPD.
It is very well known that COPD is one of the most important risk factors for the development of VAP (relative risk = 3), and VAP results in a longer ICU stay from 4 to 8 days in general ICU patient populations . Despite these data, to our knowledge, there is no study showing the incidence of VAP in patients with COPD and the effect of VAP on the duration of MV in these patients. Patients with COPD are prone to PMV because of impaired lung mechanics and respiratory muscle dysfunction. Additionally, structural abnormalities in their lungs, frequent association with bronchiectasis, repeated hospital admissions and use of antibiotics predispose these patients to colonization and infections with resistant microorganisms [17,18,19]. Due to all these reasons it was not surprising that VAP was an independent risk factor for MV >7 and >15 days in our study. But we had to include this parameter in logistic regression analysis because the objective of our study was to determine independent risk factors for PMV in patients with COPD. On the other hand, although MV >3 weeks was still significantly associated with VAP in univariate analysis it was not an independent predictor for PMV any longer. In a large series of mechanically ventilated patients, Cook et al.  showed that, although the cumulative risk for developing VAP increased over time, the daily hazard rate decreased after day 5. The risk per day was evaluated at 3% on day 5, 2% on day 10 and 1% on day 15. Our patients developed VAP on day 7. This may explain the effect of VAP on the length of MV in our study.
This study showed that older and severely ill COPD patients on admission were at high risk of a PMV >3 weeks. In previous studies, an association between high APACHE II scores and PMV has been reported. In a study by Afessa et al. , APACHE II scores correlated better than the other weaning indexes with 3- and 7-day weaning outcome in their weaning center. Their patients stayed in the ICU on average 9–11 days before weaning was initiated. In a study by Vuagnat et al. [ 22,] PMV (>14 days) was primarily determined by the diagnosis on admission and the degree of physiological impairment as measured by APACHE II score. Another study revealed that the acute physiology score and APACHE III from the 7th day of MV can be used to stratify the risk of PMV in surgical patients . However, these studies were performed in mixed ICU patients, and the characteristics of COPD patients were not studied systematically, with some studies including only a limited number of COPD patients.
A number of investigators have tried to determine if age significantly affected the duration of MV and ICU mortality. Although previous studies concluded that age had an important effect on the outcome of mechanically ventilated patients, recent research concluded that age had no effect in a mixed ICU population [24, 25]. In our study in patients with COPD, it was an independent predictor of MV for >21 days.
In this study, the degree of airflow limitation was not significantly associated with the duration of MV even in univariate analysis. Menzies et al.  found a significant relationship between the duration of MV and FEV1 in their study. In our patient group, 66% had PFT and 80% had severe and very severe airflow limitation according to the GOLD classification, with a mean FEV1 of 38% of predicted. This may explain the lack of a significant difference. Except for PaO2/FiO2 on admission, there was no significant association between blood gas values on admission and the duration of MV. PaO2/FiO2 was significantly associated with MV >7 days in this study population, but it was not an independent predictor.
Comorbidity was significantly associated with MV >21 days but did not reach the significance level as an independent predictor. Menzies et al.  could also not find a significant relationship between weaning outcome and comorbidity.
Length of MV for >7, 15 and 21 days in COPD were reported by Afessa et al. [ 21] as 57%, by Menzies et al.  as 41% and by Nevins et al.  as 9%, respectively, and the mean duration of MV ranged from 3 to 23 days in the literature [5, 6, 20]. Our results were comparable with theirs except for the lower first-week extubation rates. However, compared to these studies, in our study all the patients had COPD, their airflow limitation was more severe, they were older, had higher APACHE II scores and were more oxygen dependent (65%); 37% of the patients were unresponsive to a prior NIMV trial, 16% were using home MV and 10% had been intubated because of arrest. Despite these disadvantages, the mortality rate in this study was comparable with the studies mentioned above. The mortality rates for mechanically ventilated COPD patients in the studies ranged from 11 to 47% [6, 16]. Anon et al.  evaluated prognosis in selected patients with more severe COPD (n = 22) who had used LTOT for 3 years and reported that the mean duration of MV was 36 days, mean length of ICU stay was 41 days and ICU mortality was 35%; however, they had no control group.
Although factors influencing the duration of MV have been previously reported in different patient populations, this study differs in some important aspects. It includes only patients with COPD, assessed a wide spectrum of clinical and laboratory parameters, and tried to find the determinants of PMV for >1, 2 and 3 weeks. Results showed that the frequently associated diagnosis of CAP and bronchiectasis, and use of LTOT had no significant negative effect on MV duration in patients with COPD, but age and severity of disease on admission were the determinants of MV for >3 weeks, and VAP determined the duration of MV in the first 2 weeks.
The limitations of this study are: (1) it would be better to have a larger patient cohort for the statistical analysis in the study, and (2) this investigation describes the outcome of a single center, so our findings may not be applicable to other centers. Other factors such as strategies for weaning and sedation, infections other than VAP and other organ failures may change the percentages of patients extubated during the 1st, 2nd and 3rd week and duration of MV [27, 28.]
In conclusion, these results showed that the main factors determining MV longer than 1 or 2 weeks were the development of VAP, which was the only relatively preventable risk factor. On the other hand, MV >21 days was dependent on advanced age and severity of disease on admission.
Gul Gursel, MD
Bogaz Street 6/4
TR–06700 GOP Ankara (Turkey)
Tel. +90 312 4679344, Fax +90 312 2129019, E-Mail email@example.com
Received: February 11, 2004
Accepted after revision: June 25, 2004
Number of Print Pages : 7
Number of Figures : 0, Number of Tables : 6, Number of References : 28
Respiration (International Journal of Thoracic Medicine)
Vol. 72, No. 1, Year 2005 (Cover Date: January-February 2005)
Journal Editor: Bolliger, C.T. (Cape Town)
ISSN: 0025–7931 (print), 1423–0356 (Online)
For additional information: http://www.karger.com/res