(Chest. 2005;128:560-566.)
© 2005
American College of Chest Physicians
The Drive to Survive*
Unplanned Extubation in the ICU
James S. Krinsley, MD, FCCP and
James E. Barone, MD
* From the Critical Care Unit (Dr. Krinsley), Stamford Hospital, Stamford, CT; and Department of Surgery (Dr. Barone), Columbia University College of Physicians and Surgeons, New York, NY.
Correspondence to: James S. Krinsley, MD, Stamford Hospital, 190 West Broad St, Stamford, CT 06902; e-mail: jkrinsley{at}stamhealth.org
 |
Abstract
|
|---|
Study objectives: To assess the consequences of unplanned extubation (UE) in the ICU.
Design: Case-control study.
Setting: Fourteen-bed, medical-surgical ICU of a university-affiliated community teaching hospital.
Patients: One hundred patients who underwent UE compared to 200 control patients who underwent mechanical ventilation (MV) without UE between January 1, 1999, and June 30, 2004.
Interventions: None.
Measurements and results: Patients with UE had longer ICU and hospital length of stay (LOS) and longer duration of MV than did control subjects. Hospital mortality was 20% among UE and 35% among control patients (p = 0.011). Of the 100 patients with UE, reintubation within 48 h (UE R+) was required in 44 patients and no reintubation within 48 h (UE R–) was required in 56 patients. ICU and hospital LOS; duration of MV; rate of ICU-acquired infections; ICU pharmacy, laboratory and diagnostic imaging charges; and mortality were all much higher among UE R+ patients than among UE R– patients. Multiple logistic regression analysis revealed that age was the only predictor of the need for reintubation after UE and that age and the need for reintubation were the only predictors of mortality after UE.
Conclusions: UE was associated with increased hospital and ICU LOS but decreased mortality in this heterogeneous population of critically ill adult patients. These findings were entirely explained by the divergent outcomes of the UE R+ and UE R– groups. Patients with UE who did not require reintubation had remarkably good outcomes. It remains incumbent on ICU teams to institute protocols for regular identification of patients ready to be liberated from MV.
Key Words: ICU • intubation • mechanical ventilation • unplanned extubation
|
Introduction
|
|---|
Unplanned extubation (UE) of patients during mechanical ventilation (MV) has been reported to occur in 1 to 14% of patients, with varying rates of short-term and long-term mortality.1234567891011121314 Much of the literature regarding UE focuses on risk factors for UE or predictors of the need for reintubation after UE. Relatively few articles discuss the long-term outcomes of patients after UE; of these studies, even fewer have addressed the issue in a case-controlled manner. The studies mentioning outcomes have reported that UE has a negative or no impact on survival. This study was undertaken in order to review the outcomes of patients who have undergone UE compared to those who have not at our institution.
The case-control method was selected for the investigation, as it affords the best evidence-based medicine approach short of a randomized prospective study, which would, of course, be impossible to utilize in the setting of UE. To our knowledge, only two prior articles12 on this subject have been case-control studies.
|
Materials and Methods
|
|---|
Unit Structure and Organization
Stamford Hospital ICU is a 14-bed unit treating general medical, surgical, and cardiac patients; cardiac surgery is not performed at this institution. The hospital serves as a major teaching affiliate of Columbia University College of Physicians and Surgeons, and medical and surgical house staff as well as full-time, on-site respiratory therapists staff the unit. A 1:2 nurse:patient ratio is maintained. Although any credentialed attending physician may admit patients to the ICU, the house staff write all orders and they are closely supervised by the Director of Critical Care and the Director of Surgery, who has the certificate of Added Qualification in Surgical Critical Care. All patients receive consultations from one of the five medical intensivists or two surgical intensivists, or the appropriate medical subspecialist. Care in the unit is highly protocol driven. The unit, therefore, can best be described as "open" with a high degree of centralized control.
Critical Care Database
A detailed database of information regarding all patients admitted to the ICU has been maintained since 1998 and provides the basis for this article. The database prospectively tracks demographic, acuity, and outcome parameters of each patient admitted to the unit and is linked to numerous hospital data repositories by queries; the Director of Critical Care or his associate updates the core data set daily.
Study Design
A retrospective, case-control study was undertaken. The study group consisted of 100 patients admitted consecutively to the ICU between June 27, 1999, and September 3, 2003, who underwent UE.
The control group consisted of 200 patients. The ICU database was screened to find two control matches for each patient in the cohort group. The control patients were matched for gender, age ± 5 years, and acute physiology and chronic health evaluation (APACHE) II ± 5 points. Ninety-eight of the 100 patients with UE had two matching control patients with the same admitting service; two patients with UE matched with one patient having the same admitting service and the other with a different admitting service. UE was defined as self-extubation or accidental extubation. It was not possible to ascertain from the database whether the UE event was self-extubation or accidental, whether the patient was sedated at the time, or whether patients were being actively weaned or not at the time of extubation.
In the UE group, two groups of patients were identified: patients in whom reintubation within 48 h (UE R+) was required and patients in whom no reintubation within 48 h (UE R–) was required. The database was analyzed for prevalence of comorbidities. They were defined using the following International Classification of Diseases, Ninth Revision codes: cancer (140.00–239.99); ischemic heart disease (410.00–414.99); diabetes (250.00–250.99); respiratory disease (490.00–496.99 or 500.00–508.99); hypertension (401.99–405.99); alcoholism (303.00–303.99 or 305.00–305.99); and cerebrovascular disease (430.00–438.99). The Stamford Hospital Institutional Review Board approved this study.
Statistical Analysis
Continuous and ranked variables, including age, APACHE II scores, and LOS data were analyzed using the Wilcoxon rank-sum test for populations without normal distribution. Results were reported as median and interquartile range. Categorical variables, including gender, race, and mortality, were analyzed using the 
2 statistic. A p value of < 0.05 was considered statistically significant. All results are two tailed.
|
Results
|
|---|
There were 1,515 separate episodes of MV among 1,473 patients during the period of this study; the 100 patients, therefore, reflect a 6.6% UE rate. The hospital admission dates of the patients were distributed evenly: 47% of the patients with UE and 44% of the control patients were admitted before January 1, 2001 (p = 0.712). Therefore, a change in practice patterns over time, resulting in a change in ICU performance (ie, the "Hawthorne effect"), cannot be invoked as an explanation of the findings. The median age of the entire population of patients undergoing MV during the study period was 72 years (interquartile range, 57 to 80 years), APACHE II score was 22 (interquartile range, 16 to 28 years), and hospital mortality rate was 39.1%.
Table 1
describes demographic profiles and comorbidities of the control and UE cohort groups. The populations were very well matched for age, gender, APACHE II score, and admitting service, and there were no significant differences in the prevalence of the various underlying comorbidities.
Of the 100 patients with UE, there were 44 UE R+ patients and 56 UE R– patients. Table 2
compares the characteristics of these two populations. Multiple stepwise logistic regression revealed that age was the only predictor of the need for reintubation (p = 0.037); gender, race, admitting service, prevalence of comorbidities, and APACHE II scores were excluded from the model. Moreover, there were no statistically significant differences between the UE R+ and UE R– groups in regards to blood glucose, albumin, or hemoglobin levels at the time of UE.
Outcomes for the 100 patients with UE and their matched controls are shown in Table 3
. The patients with UE had longer ICU and hospital LOS as well as longer duration of MV. However, the ICU and hospital mortality of the patients with UE was significantly lower than that of the matched control patients. These differences are explained by the divergent outcomes of the UE R– and UE R+ groups (Table 4
). While there was no difference in the timing of UE (duration of MV prior to the UE), the UE R+ group had markedly longer ICU LOS, hospital LOS, and days of MV than did the UE R– group. There was a fivefold increase in hospital mortality among these patients and a significant difference in the number of patients acquiring infections in the ICU. Multiple stepwise logistic regression analysis, including age, APACHE II score, gender, admitting service, race, and the need for reintubation in the model, demonstrated that only age (p < 0.001) and the need for reintubation (p = 0.004) predicted hospital survival.
View this table:
[in this window]
[in a new window]
|
Table 4. Patients Requiring Reintubation Compared to Patients Not Requiring Reintubation After UE: Comparison of Selected Outcomes*
|
|
The longer ICU stay associated with UE had a large impact on resource utilization, as reflected in ICU pharmacy, imaging, and laboratory charges (Table 5
). This difference is entirely explained by the large difference in charges between the UE R+ and UE R– groups (Table 6
). Although the hospital mortality varied widely among the different groups compared (UE vs control patients, UE R+ vs UE R– patients), the discharge status of the survivors was similar (Table 7
).
View this table:
[in this window]
[in a new window]
|
Table 6. Patients Requiring Reintubation Compared to Patients Not Requiring Reintubation After UE: Comparison of Selected ICU Charges (in US Dollars)*
|
|
Finally, a comparison between the UE R– and control patients revealed no significant differences in age, gender, race, admitting service, and APACHE II scores (Table 8
). Moreover, ICU and hospital LOS, duration of MV, and laboratory, pharmacy, and imaging charges were not significantly different between the two groups. However, there was a substantial difference in hospital mortality: 7.1% among the UE R– group and 35.0% among the control subjects (p < 0.001). In comparison, the ICU and hospital mortality of the UE R+ group was similar to that of the control subjects.
|
Discussion
|
|---|
This case-control study of UE, the largest such series in the medical literature, demonstrates that patients who sustained UE were more likely to survive to hospital discharge than were those patients who did not. The survival advantage was attributable entirely to the UE R– patients. The UE R+ patients had a dramatically higher mortality rate, ICU length of stay, occurrence of ICU-acquired infections, and rate of resource utilization than did the UE R– patients.
Two case-control studies of UE have reported mortality data. Epstein et al1 performed a case-control study of 75 patients with UE compared to 150 control subjects. The 75 patients with UE represented 11% of all patients requiring MV in the ICU of their institution. The patients were matched for APACHE II score, presence of selected comorbidities, age, indication for MV, and gender. The patients with UE had longer ICU and hospital LOS. The authors described differences between the UE R+ and UE R– groups. The UE R– patients had similar mortality, ICU, and hospital LOS, and duration of MV as their control counterparts. Hospital mortality was similar among the UE patients overall compared to control subjects (32% vs 30%), and the UE R– patients did not have a statistically significant different mortality rate than did the UE R+ patients (21% vs 40%, p = 0.13).
Atkins and coworkers2 reviewed selected outcomes of 50 patients with UE compared to 100 matched control subjects. The two groups had similar age, gender, and race; UE patients were more likely to be medical rather than surgical patients. The patients with UE were more likely to have a hospital-acquired infection prior to the UE event than were the control subjects; the authors did not report the rate of infections acquired after the UE event. Hospital and ICU LOS were longer among the UE patients. Finally, the UE patients did not differ in hospital mortality compared to the control subjects (38% vs 25%, p = 0.14).
The lack of a need for reintubation after UE may identify a group of patients whose planned extubation has been delayed. Neither Epstein et al1 nor Atkins et al2 reported day of UE; the median day of UE in the current study was 3 (interquartile range, 1 to 6). Compared to the patients in the current study, the cohort of Epstein et al1 cohort was younger (mean ages, 59 years and 57 years for UE and control subjects, respectively, vs 65 years and 65 years) and had a lower APACHE II score (mean, 15 vs 20). Despite this, 56% of the UE patients in the study by Epstein et al1 study required reintubation, compared to 44% in the current study. Atkins et al2 did not compare characteristics of the UE R+ and UE R– groups in their study. However, fully 74% of their UE patients required reintubation and 51% died, compared to 0% of the UE R– patients. This high rate of reintubation after UE accounted for the trend toward higher mortality in the UE group compared to their matched controls, a finding in contrast to the mortality outcomes seen in the current study.
Other investigators have reviewed outcomes following UE using an uncontrolled case series format; Table 9
reports the salient features of these studies. The rate of reintubation after UE varied widely, from 22.6 to 74.0%, with lower rates among series dominated by surgical patients. The need for reintubation after UE was associated with higher hospital mortality rates in several series. Not surprisingly, lack of sedation or the presence of agitation predisposed to UE. A greater burden of respiratory illness, reflected by more severely impaired indexes of oxygenation, predicted the need for reintubation after UE in two studies,49 and UE R+ patients had higher APACHE II scores than did UE R– patients in two other studies.811
An important strength of our article is the large number of UE patients evaluated. Furthermore, each UE patient was matched closely with two corresponding control patients, creating statistically equivalent groups with regards to age, gender, admitting service, and APACHE II scores, as well as important comorbidities and race. The case-control methodology is the most powerful tool available to provide outcome analysis and analysis of risk factors in clinical contexts that do not allow randomized, controlled prospective trials. The comprehensive nature of the ICU database used for this study allowed detailed comparisons between the UE and control populations, as well as between the UE R+ and UE R– groups. Additional strengths of this study include analysis of charge data, rate of ICU-acquired infections, and final discharge status of each patient.
The main limitation of this study is the absence of data relating to the circumstances surrounding episodes of UE. Bedside records were not reviewed to abstract information about the use of restraints, dosage of sedation, size of endotracheal tubes, whether the UE episode was "accidental" or "deliberate," and other "soft" clinical data points.
Based on our article and the recent literature, some general observations can be made. The reported rate of UE in most ICUs is approximately 7 to 10%. Approximately one half of all UE patients will require reintubation. Patients who undergo UE and do not require reintubation may actually have better outcomes than do patients receiving MV without UE. This may be due to the acceleration of the weaning process in those not requiring reintubation, or it may simply be that those who tolerate UE are physiologically better equipped to survive, despite the lack of differences in APACHE II scores or comorbidities.
The 7% hospital mortality rate of our UE R– group is remarkable, considering their baseline median APACHE II score of 20. The UE R+ group had a 35% mortality rate, which is similar to that of the control group. In this study, as well as the other two case-control studies,12 the difference in mortality between the UE and control patients was driven by the differences in the rate of reintubation after UE. It is difficult, if not impossible, to predict who will not need reintubation after UE, except perhaps that young patients as a group may fare better. Indeed, no other demographic or clinical factor differentiated the UE R+ and UE R– groups in this study. While it may be possible to reduce the incidence of UE using the performance improvement process, it remains incumbent on ICU staffs to institute protocols for regular identification of patients ready to be liberated from MV.
|
Footnotes
|
|---|
Abbreviations: APACHE = acute physiology and chronic health evaluation; LOS = length of stay; MV = mechanical ventilation; UE = unplanned extubation; UE R+ = reintubation within 48 h; UE R– = no reintubation within 48 h
Received for publication September 15, 2004.
Accepted for publication January 5, 2005.
|
References
|
|---|
- Epstein, SK, Nevins, ML, Chung, J (2000) Effect of unplanned extubation on outcome of mechanical ventilation. Am J Respir Crit Care Med 161,1912-1916[Abstract/Free Full Text]
- Atkins, PM, Mion, LC, Mendelson, W, et al Characteristics and outcomes of patients who self-extubate from ventilatory support: a case-control study. Chest 1997;112,1317-1323[Abstract/Free Full Text]
- Moons, P, Sels, K, De Becker, W, et al Development of a risk assessment tool for deliberate self-extubation in intensive care patients. Intensive Care Med 2004;30,1348-1355[Medline]
- Chevron, V, Menard, JF, Richard, JC, et al Unplanned extubation: risk factors of development and predictive criteria for reintubation. Crit Care Med 1998;26,1049-1053[CrossRef][ISI][Medline]
- Razek, T, Gracias, V, Sullivan, D, et al Assessing the need for reintubation: a prospective evaluation of unplanned endotracheal extubation. J Trauma 2000;48,466-469[Medline]
- Betbese, AJ, Perez, M, Bak, E, et al A prospective study of unplanned endotracheal extubation in intensive care unit patients. Crit Care Med 1998;26,1180-1186[CrossRef][ISI][Medline]
- Boulain, T Unplanned extubations in the adult intensive care unit: a prospective multicenter study. Am J Respir Crit Care Med 1998;157,1131-1137[Abstract/Free Full Text]
- Phoa, LL, Pek, WY, Syap, W, et al Unplanned extubation: a local experience. Singapore Med J 2002;43,504-508[Medline]
- Whelan, J, Simpson, SQ, Levy, H Unplanned extubation: predictors of successful termination of mechanical ventilatory support. Chest 1994;105,1808-1812[Abstract/Free Full Text]
- Coppolo, DP, May, JJ Self-extubations: a 12-month experience. Chest 1990;98,165-169[Abstract/Free Full Text]
- Tindol, GA, Jr, DiBenedetto, RJ, Kosciuk, L Unplanned extubations. Chest 1994;105,1804-1807[Abstract/Free Full Text]
- Pandey, CK, Singh, N, Srivastava, K, et al Self-extubation in intensive care and re-intubation predictors: a retrospective study. J Indian Med Assoc 2002;100,14-16
- Tominaga, GT, Rudzwick, H, Scannell, G, et al Decreasing unplanned extubations in the surgical intensive care unit. Am J Surg 1995;170,586-589[CrossRef][ISI][Medline]
- Maguire, GP, DeLorenzo, LJ, Moggio, RA Unplanned extubation in the intensive care unit: a quality-of-care concern. Crit Care Nurs Q 1994;17,40-47[Medline]
TOP
Abstract
Introduction
