American Journal of Respiratory and Critical Care Medicine

The predisposing factors and complications of unplanned extubation (UEX) in mechanically ventilated adult patients are not well recognized. We designed a prospective multicenter observational study to identify risk factors and describe the complications of UEX. We followed 426 ventilated patients over a 2-mo period. Clinical characteristics such as diagnosis on admission and reasons for ventilation were used to classify the patients. The presence or absence of potential risk factors was daily noted, including the types of ventilators, tracheal tubes, tube fixations, ventilatory support modes, route for intubation, and the use of intravenous sedation. Circumstances and complications of UEX were prospectively recorded. Forty-six (10.8%) patients experienced at least one episode of UEX. Ten UEX occurred during nursing procedures. At the moment of UEX, 61% of patients were agitated. The rates of mortality, laryngeal complications, nosocomial pneumonia after extubation, and the length of mechanical ventilation were similar in UEX and non-UEX patients. Patients were more often reintubated after UEX (28 of 46) than after planned extubation (28 of 284). All the non-reintubated UEX patients survived. One death occurred as a direct consequence of UEX. By use of multivariate analysis, we identified four factors contributing to UEX: chronic respiratory failure, endotracheal tube fixation with only thin adhesive tape, orotracheal intubation, and the lack of intravenous sedation. Considering these factors, we hypothesized that simple measures should be adopted to minimize the incidence of UEX and its related complications: more vigilance during procedures at patients' bedsides, adequate sedation of agitated patients, strong fixation of the tracheal tube, particular attention paid to orally intubated patients, and daily reassessment of the possibility of weaning from the ventilator.

Among complications of endotracheal intubation for respiratory support in critically ill adult patients, unplanned (i.e., self or accidental) extubation (UEX) is a frequent event (1) that can cause morbidity and mortality (2). Factors contributing to this event are not well recognized in adult patients, and it is unclear, for instance, whether factors such as the route of tracheal intubation, the method of tube fixation, or the degree of patient's consciousness could predispose patients to UEX, as studies with conflicting results have been published (3, 4). Furthermore, although life-threatening complications of UEX have been described by several authors (2, 5-7), other possible consequences of UEX such as pneumonia or additional length of mechanical ventilation (MV) have not been systematically examined. We therefore designed a prospective multicenter study intended to define the predisposing factors, to assess the complications, and to suggest means of preventing UEX.

Eleven medical or medical-surgical intensive care units (ICUs) for adult patients in 11 hospitals (three university hospitals, one regional and teaching hospital, and seven general hospitals) participated in the study. All of the patients admitted and intubated in the 11 ICUs within the 2-mo period from April 1, 1992, to May 31, 1992, were included. Patients who were tracheotomized before admission were excluded.

Data Collection

Prospective data collection included three sections:

1. Usual clinical characteristics such as age, sex, severity acute physiology score (SAPS) (8), diagnosis on admission, MV duration, and outcome were recorded for each patient.

2. Daily monitoring was conducted at the bedside by the attending physician, who noted numerous clinical and therapeutical data such as tracheal tube type and fixation used, ventilatory support mode, use of intravenous sedation or not, oro- or nasotracheal route of intubation, and patients' mobility and position.

3. Every time a UEX event occurred, a specific questionnaire was immediately filled in by the attending physician. Time, date, ventilatory support mode, degree of consciousness, nurse/patient ratio, and particular circumstances were among the data recorded.

Complications of UEX were also recorded: respiratory distress, need for reintubation, bradycardia, laryngeal complications, and nosocomial pneumonia. These last two possible complications were also prospectively recorded in non-UEX patients.

Because of the noninterventional nature of our study, ethical approval and patients' consent were judged unnecessary.

Definitions

We defined UEX as patient's extubation unwanted by the medical staff. Thus, UEX could be extubation by the patient himself (deliberately or accidentally) or accidental extubation by members of the auxiliary nursing and medical staffs during procedures at the bedside.

Laryngeal complications of extubation were defined as (1) clinically evident laryngeal obstructive dyspnea occurring within 24 h after UEX or after planned extubation, and (2) oro- or nasopharyngeal bleeding within 24 h after extubation, requiring hemostatic intervention or red-blood-cell transfusions.

Nosocomial pneumonia was affirmed when chest radiograph showed new pulmonary infiltrates 48 h after admission, not otherwise explained, in combination with at least two of the following symptoms: (1) fever greater than 38° C, (2) purulent tracheobronchial secretions, and (3) peripheral leukocyte count greater than 10,000/mm3.

Variables Studied as Potential Risk Factors for Unplanned Extubation

The ICU of hospitalization, sex, age, SAPS, presence or absence of coma by drug overdose, chronic respiratory failure (CRF), acute exacerbation of CRF, chronic alcoholism, previous episode of MV for more than 24 h, previously known psychiatric disease or encephalopathy, and the length of MV were recorded for all patients and tested as categorical risk factors for UEX.

The patients were classified in medical, surgical, or traumatic categories, which were studied as potential risk factors for UEX. Patients were considered to be trauma patients, regardless of whether they needed surgery or not, when a traumatic accident caused their hospitalization. All patients who underwent surgery between 24 h before and 48 h after admission to the ICU were classified as surgical patients (trauma patients excluded). The remainder were considered medical patients.

The route of intubation (oro- or nasotracheal), the number of intubations, the endotracheal tube type (manufacturer, low-pressure cuff or not), the type of tube fixation used (“lasso,” strong or thin adhesive tape), tube diameter, respiratory support mode (five different modes), type of ventilator (10 types), and the use of intravenous sedation were noted each day at the bedside. All these categorical variables were studied as potential risk factors for UEX.

The respiratory support modes used were the following: volume-controlled mode (VC), intermittent mandatory ventilation, continuous positive airway pressure ventilation, pressure-support ventilation, and spontaneous ventilation through the tracheal tube with no mechanical ventilator. Two patients who had been ventilated for several days in the pressure-controlled mode were recorded as being ventilated in the VC mode. The methods of securing tracheal tubes were classified in three different types (for data analysis, each type was recorded as present [0] or absent [1]): (1) thin, 1-cm adhesive tape wrapped around the tube and applied on both cheekbones, regardless of whether it circled the neck or not; (2) strong, 2-cm cloth adhesive tape, regardless of whether it circled the neck or not; or (3) cloth string wrapped around the tube and circling the neck. Bite blocks were not used. The ventilators used were the following: Servo 900 A, B, C, and D (Siemens, Solna, Sweden), 7200a (Puritan-Bennett, Carlsbad, CA), César (CFPO, Paris, France), CPU1 (Ohmeda, Trappe, France), Evita (Dräger, Lübeck, Germany), Engstrom 150 (LKB Medical-Engstrom, Stockholm, Sweden) and Monnal M1 (CFPO). The endotracheal tubes used were the following: Portex Blue-line (LSSA, Fontenay-sous-Bois, France), Mallinckrodt Lo-Pro (Medexel, Hennes/Sieg, Germany), ORX (Vygon, Ecouen, France), Aire-Cuf (Bivona, Gary, IN), and Ruschlit (Rusch, Waı̈bligen, Germany).

Statistical Analysis

Quantitative values were expressed as mean ± SD and compared between UEX and non-UEX patients using Student's t test. Qualitative data were expressed as percentages and compared by chi-square test. A p value < 0.05 was considered significant.

The probability of UEX occurrence according to the presence or absence of variables expected as risk factors was estimated by several means. First, the odds ratios (OR) were used to identify variables that could not change throughout the study period, such as sex, chronic alcoholism, CRF, admission for trauma, etc., as risk factors; their statistical significance was ascertained by a 95% confidence interval (95% CI). Then, after reducing the number of variables to those that met the 0.2 significance level, the ORs of each variable were adjusted for the remaining variables (adjusted ORMH) according to Mantel and Haenszel (9).

Second, the daily probability of UEX occurrence over time, according to the presence or absence of potential risk factors, was calculated using the Kaplan-Meier estimate and the log-rank test (10). Covariates that met a significance level of 0.05 with this univariate analysis were then adjusted for the other baseline characteristics using a Cox proportional-hazards model (11). For this day-to-day analysis, the study population had to be reduced to a subset of patients for two reasons: (1) the stratified table constructed disclosed some empty cells due to the observational, rather than controlled, nature of the study, and (2) patients' conditions could change during the study (e.g., some patients could be orally and then nasally intubated).

Third, in an attempt to consider factors that could alternately be present or absent during the ICU stay of each patient, we also calculated the percentage of days with UEX compared to the total number of intubated days and the presence or absence of each potential risk factor. Crude ORs and adjusted ORMH were calculated. This approach, although difficult to interpret, was chosen because time-dependent multivariate analysis could not be conducted since numerous patients had changing potential risk factors. For example, patients were often alternately ventilated in volume-controlled and pressure-support modes, or alternated days with sedation and days without sedation. Moreover, independently of the changes in ventilation or sedation modes, patients could, for instance, alternate different routes for intubation throughout the study.

Log-rank tests and the Cox proportional-hazards model were performed with the StatView 4.02 statistical package (Abacus Concepts Inc., Calabasas, CA). Adjusted ORMH was calculated with the Epi-Info v.5 package (Centers for Disease Control, Atlanta, GA).

Population Characteristics

A total of 426 ventilated patients were admitted to the ICUs during the 2-mo study period. Demographic and clinical characteristics are set out in Table 1. No patient was admitted more than once during this period. All patients were included and prospectively followed until discharge from the ICU or until death. The percentage of patients included in each ICU was not significantly different from the percentage expected, according to the contribution of each ICU to the total number of beds. The three university ICUs and the regional hospital together contributed nearly 50% of the study population.

Table 1. DEMOGRAPHIC AND CLINICAL CHARACTERISTICS OF THE POPULATION STUDIED (n  =  426)

Sex ratio, M/F262/164
Age, yr  57 ± 19 (12–94)
Admission SAPS13.9 ± 5.3 (1–34)
Days under MV 8.4 ± 8.7 (1–52)
Days under MV except days after UEX or tracheotomy*  7.8 ± 8.4 (1–38)
Survivors321 (75.4%)
Surgical patients64 (15%)
Trauma patients36 (8.5%)
Medical patients326 (76.5%)
Drug overdose82 (19%)
CRF92 (22%)
Acute exacerbation of CRF60 (14%)
Alcohol addiction79 (18%)
Previous episode of MV82 (19%)
Orotracheal intubation145 (34%)
Nasal intubation241 (57%)
Oral and nasal intubation during ICU stay40 (9%)

Definition of abbreviations: SAPS = severity acute physiology score; MV = mechanical ventilation; UEX = unplanned extubation; CRF = chronic respiratory failure. Data are given as mean ± SD (range) or as number (percent).

* Thirty-one patients (7%) underwent a tracheotomy during their ICU stay.

Overall, the surveillance represented 3,576 patient-ventilated days, or a total of 3,312 patient-ventilated days after excluding ventilated days after UEX or after a tracheotomy had been performed.

Characteristics of the Unplanned Extubation Population

Of the 426 study patients, 46 (10.8%) experienced at least one episode of UEX. Among them, five patients had two episodes each and two patients had four episodes each. As a result, we observed a total of 57 UEX, corresponding to a rate of 1.59 UEX per 100 intubated days.

Unplanned extubation patients were ventilated for 5.2 ± 8.1 d (range, 1–38 d) before their first episode of UEX, and their total MV duration averaged 10.9 ± 10.6 d (range, 1–42 d). Their mean age was 60 ± 21 yr. Their admission SAPS were 13 ± 3.2, not different from those of non-UEX patients. Thirty patients were orally intubated and 27 nasally intubated at the moment of UEX.

At the time of UEX, 46% of patients (26 of 57) were considered cooperative, 61% (35 of 57) were agitated, 12% (7 of 57) were receiving continuous intravenous sedation, and 58% (33 of 57) had their hands restrained. Fifty-five times, UEX was considered to be self-extubation, deliberate or not (including two UEX partially due to an air leak in the tube cuff and two UEX probably due to tube position). Twice, UEX was accidentally caused by a member of the nursing staff. Overall, 10 UEX occurred during nursing procedures (washing or changing the patient) at the bedside. The number of UEX was evenly distributed among the different days of the week and among the different times of day. The nurse/patient ratio at the moment of UEX was not different from the mean ratio generally observed in each ICU (ranging from 1 of 3 to 1 of 4).

Risk Factors

Among baseline clinical characteristics that could not change during the study period, only the presence of CRF was identified as a risk factor for UEX (Table 2). The facility had no influence on the probability of UEX.

Table 2. UNPLANNED EXTUBATION OCCURRENCE BY CLINICAL CATEGORIES

CategoriesNon-UEX PatientsUEX PatientsCrude Odds Ratio (95% CI)Adjusted Odds RatioMH *(95% CI)
Male patients
 Yes23230 1.2 (0.5–2.9)
 No14816
Surgical patients
 Yes 57 71.02 (0.84–1.22)
 No32339
Trauma patients
 Yes 29 72.17 (0.77–6.12)
 No35139
Medical patients
 Yes294320.67 (0.3–1.48)
 No 8614
Drug overdose
 Yes 76 6 0.6 (0.2–1.75)
 No30440
CRF
 Yes 76162.13 (1.06–4.3)2.1 (1.5–3)
 No30430
Acute exacerbation of CRF
 Yes 50101.83 (0.76–4.4)
 No33036
Alcohol addiction
 Yes 66131.87 (0.87–4.06)
 No31433
Previous episode of MV
 Yes 77 52.08 (0.7–6.2)
 No30341
Orotracheal intubation only
 Yes128171.15 (0.42–3.18)
 No25229
Nasal intubation only
 Yes218260.97 (0.21–4.4)
 No16220
Oral and nasal intubation before UEX
 Yes 34 3 0.7 (0.06–8.24)
 No34643
ICU #4
 Yes 16 63.41 (1.15–10.2)
 No36440

For definition of abbreviations, see Table 1.

* OR is shown only if it reaches the p = 0.2 significance level.

After patients who alternated methods for tube fixation, routes of intubation, tube types, or ventilators during their period of intubation and all 38 patients from one center where no UEX occurred had been excluded, the study population was composed of 380 patients, including 38 UEX patients. In this restricted population, the day-to-day univariate analysis disclosed three risk factors for UEX: lack of strong tube fixation (i.e., fixation with only thin adhesive tape; log rank = 4.15; p = 0.041), orotracheal intubation (log rank = 4.07; p = 0.043), and presence of CRF (log rank = 3.95; p = 0.047). After adjustment using the Cox proportional-hazards model, orotracheal intubation (OR = 3.6; 95% CI, 1.26–10.2) and CRF (OR = 2.43; 95% CI, 1.17–5.07) remained as independent risk factors for UEX.

Considering intubated days with the presence or absence of potential risk factors, seven factors appeared as independent risk factors (Table 3). Three of those were found to be significant: lack of strong tube fixation, lack of intravenous sedation, and orotracheal intubation. Despite the lack of logistic analysis, it could easily be observed that these three factors interacted in an additive or multiplicative manner: the UEX rate when no factor or only one was present ranged from 0 to 0.5 per 100 ventilated days, ranged from 2.3 to 3.6 per 100 ventilated days with two factors, and reached the rate of 4.1 per 100 ventilated days when all three factors were present.

Table 3. UNPLANNED EXTUBATION OCCURRENCE ACCORDING TO DAY-TO-DAY EXPOSURE TO RISK FACTORS

Risk FactorNon-UEX Intubation (d )UEX Intubation (d )Crude Odds Ratio*,(95% CI)Adjusted Odds RatioMH *(95% CI)
Lack of strong tube fixation
 Yes 501274.9 (2.6–9.2)3.45 (2.3–5.2)
 No2,76519
Orotracheal intubation
 Yes 816253.6 (2.06–6.4) 3.7 (2.7–5)
 No2,49721
Lack of intravenous sedation
 Yes2,282413.5 (1.4–8.9) 5.9 (3.5–9.9)
 No 984 5
Pressure support ventilation
 Yes2,803121.9 (1.39–2.7)
 No 51034
Partial MV
 Yes 670151.88 (0.96–3.7)
 No2,61031
Volume-controlled ventilation
 Yes2,518282.04 (1.3–3.2)
 No 79518
ICU #4
 Yes 199 62.3 (0.85–6.3)
 No3,06740

For definition of abbreviations, see Table 1.

* OR is shown only if it reaches the p = 0.2 significance level.

Partial MV includes all types of ventilation modes except volume-controlled mode and spontaneous ventilation.

Complications of Unplanned Extubation

The mortality rate was similar in UEX (7 of 46) and non-UEX (98 of 380) patients. The rate of nosocomial pneumonia was similar in UEX (4 of 46) and in non-UEX patients (22 of 380); in fact, the four cases of nosocomial pneumonia in the UEX population occurred before the episode of UEX. When only the first episodes of extubation were considered for each patient, the rate of reintubation after UEX was 61% (28 of 46). This rate was higher (p < 0.0001) than the reintubation rate of 10% (28 of 284) observed in the non-UEX population after planned extubation. The UEX patients who had CRF were reintubated more frequently than non-UEX patients (14 of 16 and 13 of 30, respectively; p < 0.01). Of a total of 28 reintubations in the UEX population, 27 were performed within 72 h (a delay generally used to define successful weaning from the ventilator [12]). Among them, 19 were done within 2 h for life-threatening respiratory failure. None of the non-reintubated UEX patients died. In the non-UEX population, the attempt to wean them from the ventilator led to reintubation within 72 h in 22 cases considered to be failures of weaning. The reason for reintubation was respiratory distress in all cases.

The additional length of MV after the first extubation was similar in the reintubated UEX patients and in the reintubated non-UEX patients (9.5 ± 6 d versus 10.5 ± 7 d, respectively). This additional length of MV was also similar (by analysis of variance) between UEX and non-UEX patients in the subgroups of CRF and non-CRF patients.

The rate of immediate laryngeal edema after extubation was similar in UEX patients (2 of 46) and in non-UEX patients (12 of 284). We observed no bleeding complication of extubation in either group. One prolonged cardiac arrest, causing death 2 d later, complicated one episode of UEX in a postoperative 41-yr-old patient.

Unplanned extubation is a frequent event in the adult ICU and leads to few but potentially fatal complications. Our study, which is the first large multicenter study focused on UEX and using multivariate data analysis, strongly suggests that the use of nasal rather than oral intubation, the continuous use of sedation until weaning is begun, and the systematic use of very strong tube fixation could prevent UEX. We also hypothesize that UEX rate could be a useful indicator of nursing and medical care quality in the ICU.

Unplanned Extubation Incidence

We observed 10.8% of UEX events in our population, a rate which is similar to the rates of 7% to 16% found in other studies (1, 2, 5, 7, 13, 14). Little and colleagues (5) suggested that UEX incidence should be expressed as the number of UEX observed per 100 ventilated days to allow comparisons among different ICUs, because one can reasonably consider that the longer patients are ventilated, the longer they are exposed to the risk of UEX. Unfortunately, few authors have expressed their results in such a manner. Therefore, the apparent similarity among UEX rates reported in several studies could be false if reassessed considering ventilated days. Orlowski and colleagues (15) in children and recently Tindol and colleagues (4) in adult patients reported UEX frequencies of 4% and 3%, far below the rates reported by others. However, the lack of data expressed in terms of ventilated days makes it difficult to know if patients were accidentally extubated less often. When recalculating UEX incidence from available prospective studies reporting the length of ventilation (1, 2), we found a rate of UEX ranging from 1.1 to 2.1 per 100 ventilated days, similar to the rate of 1.6 we observed. Therefore, it could be considered that UEX occurs with rather comparable frequencies in different adult ICUs. At first glance, one might presume from these studies (1, 2) and ours that a given frequency of UEX is unavoidable in ventilated patients, particularly since UEX could not be avoided despite the staff's awareness that a study was being performed. However, we think that numerous cases of UEX could be avoided in view of the predisposing factors pointed out in our study that could be easily minimized.

Risk Factors

Factors influencing the occurrence of UEX are not numerous in the literature. Little and colleagues (5), in a large study including newborns and infants, found that procedures at the bedside such as portable roentgenogram, turning, and changing the patients put them at risk of UEX. However, this association was not examined prospectively. In our study, 10 patients experienced UEX during turning or washing procedures, a rate we consider low in view of the very large number of such procedures performed throughout the study. Clearly, however, these UEX could have been avoided.

As in Vassal and colleagues' study (2), CRF appeared as a risk factor for UEX in our study population, a result that was not found in earlier studies. This association was not explained in our study by the longer time spent under MV by CRF patients: CRF patients were ventilated 7.1 ± 9 d before UEX, and non-CRF patients were ventilated 4.2 ± 7.6 d before UEX (p = 0.24). The fact that CRF patients are generally put under partial ventilation mode early during their ICU stays may explain their tendency to inadvertantly extubate themselves. However, we failed to demonstrate any influence of partial ventilatory support on UEX when UEX probability was expressed per ventilated days. Our results tend to show that CRF patients were prone to UEX because they spent more time under MV without sedation due to their longer weaning process. Although we found CRF as the sole clinical baseline characteristic that could predispose a patient to UEX in the first step of our data analysis, CRF did not remain as a risk factor in the analysis that took intubated days into account, while the lack of intravenous sedation did so.

Obviously, the lack of sedation as a predisposing factor for UEX was not unexpected: agitated patients become less prone to UEX when movement is prevented by adequate sedation. Nevertheless, sedation is not always used to prevent patients' motion, and periods of progressive withdrawal from sedation are necessary. During these critical periods, means available to prevent accidental events like UEX in agitated patients are not numerous. Restraints on a patient's wrists and chest can be used to prevent UEX (16). In some studies, it has been suggested that restrained, agitated patients who extubated themselves should have been attached more securely (4, 16). We agree that bonds could prevent unintentional UEX. However, determined patients are often able to extubate themselves even when they are strongly restrained (1). Rather than bonds that look like a straitjacket, we think that adequate intravenous sedation should be employed for uncontrollable agitation in critical care (17). In fact, agitation in critically ill patients should be considered an alarming symptom that should prompt further investigations in search of encephalopathy of infectious, metabolic, hypoxemic, low-flow state, drug-induced, or traumatic nature. Therefore, agitated patients should not reasonably be considered candidates for weaning from the ventilator. In our opinion, attempts to wake up these patients before appropriate care has been initiated are not advisable. Sedation titrated to obtain a score of 5 on Ramsay's scale appears to be comfortable for both patients and nursing staff, as described twenty years ago (18).

We did not confirm the influence of age on UEX occurrence reported by Vassal and colleagues (2). Their population included 39% patients with CRF, compared with 21% in our study. The fact that CRF patients are generally older and spent more time intubated without sedation than other patients could explain why age was identified as a risk factor. However, firm conclusions cannot be drawn since these authors did not perform multivariate or day-to-day analysis.

The predisposing effect of oral intubation, compared to nasal intubation, on UEX frequency has been suggested by Coppolo and May (1) and by Ripoll and colleagues (3). This effect has been found significant by McMillan and colleagues (19) in critically ill children. Recently, Tindol and colleagues (4) failed to demonstrate any effect of the route of intubation, probably because their study suffered from the lack of multivariate analysis. Vassal and coworkers (2) were unable to draw a conclusion, since their study population was too small and composed mostly of nasally intubated patients. For our part, we found that orotracheal route for intubation made patients more prone to UEX compared to nasal intubation, by both multivariate analysis (OR = 3.6) and analysis that considered intubated days (OR = 3.7). This result can be easily explained by the high mobility of the endotracheal tube during patients' motions when they are orally intubated (20). Moreover, uncooperative patients are often able to move the tube with their tongues (20). Flexion or extension of the head, which occur frequently during procedures at the bedside, make the endotracheal tube move more easily with oral than with nasal intubation (3, 20). It should be noted, however, that during prolonged ventilation of the critically ill patient, oral intubation's higher risk of UEX is counterbalanced by its lower risk of nosocomial maxillary sinusitis (21).

When evaluated in terms of intubated days, motions of the endotracheal tube and UEX are more frequent during oral intubation when only fragile tube fixation is used or when patients are not sedated, both factors facilitating tube mobility (22). The effect of the type of tube fixation on UEX had not been found in previous studies (7). Systematic tracheotomy for ventilation could apparently avoid a vast number of UEX. Unfortunately, it does not seem to be true, since 7.45% (29 of 389) of tracheotomized patients were inadvertantly extubated in the study by McClelland (23).

For reasons that are probably related to different routines among ICUs, some authors found UEX more frequent in the morning (2, 4) and others in the evening (1), while UEX was unvaryingly distributed over the day in our study.

Complications of Unplanned Extubation

As observed in previous studies, we did not find an increased mortality in patients who experienced UEX. However, UEX patients are not exempt from serious complications (1). We observed one death directly related to the UEX episode, a consequence that is not rare (2, 5-7). Surprisingly, although already observed by Rashkin and colleagues (24), immediate laryngeal complications after UEX were not frequent (2 of 46), considering the traumatic nature of extubation with the tube cuff inflated. Zwillich and colleagues (14) in 1974 observed more nosocomial pneumonia in UEX patients. There is no clear explanation why we did not find this result. It could be related to differences in patient recruitment and possibly to differences in general care approaches between the 1970s and today.

Although cautious clinical evaluation can often avoid immediate reintubation of UEX patients (1, 14, 25), reintubation remains necessary in 31% (1) to 74% (2) of the patients. In our population, 28 of 46 patients (61%) were reintubated after UEX, usually because of respiratory distress. Although not observed in our study, reintubation could also have led to serious complications (24, 26, 27).

Our data concerning the rate of reintubation and the subsequent additional length of MV are difficult to interpret. Regarding the fact that 27 of 28 patients were reintubated within 72 h, it could be said that these patients were not ready for weaning from the ventilator. Indeed, their additional length of MV (a potential complication that had never been studied before) made their total length of MV similar to that for reintubated non-UEX patients. On the other hand, regarding the number of patients reintubated within a few hours (19 of 28), it could be argued that these patients were reintubated because of the stress induced by UEX (e.g., increased oxygen demand, muscle fatigue) and that their additional length of MV was undue, putting them at risk of further complications.

Among our 46 UEX patients, 18 were not reintubated and all 18 survived. These patients were ready for weaning from the ventilator, and extubation should have been attempted earlier by the medical staff in order to avoid serious complications.

Conclusion

Although MV is useful in critically ill patients even when applied for long periods (28), ventilated patients are exposed to additional complications due to MV itself (26). Among these complications, UEX is a frequent event that can lead to serious immediate complications, such as an ischemic cardiac arrest, or delayed complications related to the necessity of reintubation and perhaps to an undue additional length of MV. Our study definitively shows that a lack of intravenous sedation, the orotracheal route for intubation, and a lack of strong tube fixation are the major predisposing factors to UEX in ventilated adult patients. Although it could appear difficult to improve the efficiency of patients' surveillance and monitoring, particularly in our French ICUs where the nurse/patient ratio is rarely over 1:3, we think that more vigilance by the medical and nursing staffs might lower the incidence of UEX. On the one hand, more vigilance by nurses and auxilliary nurses during nursing procedures seems mandatory to avoid a significant number of UEXs. On the other hand, the medical staff should also adopt measures intented to protect the patients: (1) more comprehensive treatment of agitated patients; (2) verification of the correct position of the endotracheal tube daily (or more frequently) in patients who are orally intubated; (3) strong attachment of the tracheal tube, particularly in orally intubated patients; and (4) daily reconsideration of the possibility of extubation. To conclude, we think that the survey of UEX occurrence could be part of a global assessment of the quality of care, like the rates of nosocomial infections, pressure sores, or mortality, and used to compare the quality of care among ICUs.

1. Coppolo D. P., May J. J.Self-extubation: a 12-month experience. Chest981990165169
2. Vassal T., Anh N. G. D., Gabillet J. M., Guidet B., Staikowsky F., Offenstadt G.Prospective evaluation of self-extubations in a medical intensive care unit. Intensive Care Med.191993340342
3. Ripoll I., Lindholm C.-E., Carroll R., Grenvik A.Spontaneous dislocation of endotracheal tubes. Anesthesiology4919785052
4. Tindol G. A., DiBenedetto R. J., Kosciuk L.Unplanned extubations. Chest105199418041807
5. Little L. A., Koenig J. C., Newth C. J. L.Factors affecting accidental extubations in neonatal and pediatric intensive care patients. Crit. Care Med.181990163165
6. Molho M., Lieberman P.Safe fixation of oro- or nasotracheal tubes for prolonged intubation in neonates, infants and children. Crit. Care Med.319758182
7. Scott P. H., Eigen H., Moye L. A., Geogitis J., Laughlin J. J.Predictability and consequences of spontaneous extubation in a pediatric ICU. Crit. Care Med.131985228232
8. Le Gall J. R., Loirat P., Alperovitch A., Glaser P., Granthil C., Mathieu D., Mercier P., Thomas R., Villiers D.A simplified acute physiology score for ICU patients. Crit. Care Med.121984975977
9. Mantel N., Haenszel W.Statistical aspects of the analysis of data from retrospective studies of disease. J. Nat. Cancer Inst.221959719748
10. Peto R., Pike M. C., Armitage P., Breslow N. E., Cox D. R., Howard S. V., Mantel N., McPherson K., Peto J., Smith P. G.Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br. J. Cancer351977139
11. Cox D. R.Regression models and life-tables. Journal of the Royal Statistical Society341972187220
12. Gibert C., Bourgain J. L., Brunet F., Derenne J. P., Daoudal P., Manier G., Roeseler J., Thomas R.Le sevrage de la ventilation mécanique chez l'adulte. VIIIème Conférence de Consensus en Réanimation, Lille, France, Juin 1991. Réanimation Urgences11992159167
13. Jayamanne, D., R. Nandipati, and D. Patel. 1988. Self-extubation: a prospective study (abstract). Chest 94(Suppl. 1):3S.
14. Zwillich C. W., Pierson D. J., Creagh C. E., Sutton F. D., Schatz E., Petty T. L.Complications of assisted ventilation: a prospective study of 354 consecutive episodes. Am. J. Med.571974161170
15. Orlowski J. P., Ellis N. G., Amin N. P., Crumrine R. S.Complications of airway intrusion in 100 consecutive cases in a pediatric ICU. Crit. Care Med.81980324331
16. Brandstetter, R. D., I. T. Khawaja, and E. Bartky. 1991. Self-extubation. Chest 99:1319–1320. Letter.
17. Windsor D., Ward B., Albrant D., Roma M., Trask A., Rainey T.Reduction of unintentional extubations among agitated ICU patients: an algorithm for sedation (abstract). Crit. Care Med.231995A46
18. Ramsay M. A. E., Savege T. M., Simpson B. R. J., Goodwin R.Controlled sedation with alpharealone-alphadalone. B.M.J.21974656659
19. McMillan D. D., Rademaker A. W., Buchan K. A., Reid A., Machin G., Sauve R. S.Benefits of orotracheal and nasotracheal intubation in neonates requiring ventilatory assistance. Pediatrics7719863944
20. Conrardy P. A., Goodman L. R., Lainge F., Singer M. M.Alteration of endotracheal tube position: flexion and extension of the neck. Crit. Care Med.41976812
21. Salord F., Gaussorgues P., Marti-Flich J., Sirodot M., Allimant C., Lyonnet D., Robert D.Nosocomial maxillary sinusitis during mechanical ventilation: a prospective comparison of orotracheal versus the nasotracheal route for intubation. Intensive Care Med.161990390393
22. Levy H., Griego L.A comparative study of oral endotracheal tube securing methods. Chest104199315371540
23. McClelland R. M. A.Complications of tracheostomy. B.M.J.21965567569
24. Rashkin M. C., Davis T.Acute complications of endotracheal intubation: relationship to reintubation, route, urgency, and duration. Chest891986165167
25. Whelan J., Simpson S. Q., Levy H.Unplanned extubation: predictors of successful termination of mechanical ventilatory support. Chest105199418081812
26. Blanc V. F., Tremblay N. A. G.The complications of tracheal intubation: a new classification with a review of the literature. Anesth. Analg.531974202213
27. Legras, A., et l'Association des Réanimateurs du Centre Ouest. 1994. Intubation difficile en réanimation: facteurs de risque et complications. Etude prospective multicentrique (abstract). Réanimation Urgences 3:698 (SP88).
28. Gracey D. R., Naessenes J. M., Kishan I., Marsh H. M.Hospital and posthospital survival in patients mechanically ventilated for more than 29 days. Chest1011992211214
Correspondence and requests for reprints should be addressed to Dr. T. Boulain, Service de Réanimation Centre Hospitalier Régional d'Orléans-La Source, BP 6709, 45067 Orléans Cedex 2, France.

The Association des Réanimateurs du Centre-Ouest (ARCO) is composed of intensivists who work in 24 Intensive Care Units of the middle and west regions in France. The following ARCO members participated in this study. Angers: G. Bouachour, J. P. Gouello; Blois: F. Fournier, C. Hennequin; Brive: E. Karam, P. Chevallier; Chateauroux: M. Orillard, R. Parisot; Dreux: V. Julie, N. Letellier; Le Mans: N. Varache, C. Haas; Orléans: M. Wolf, C. Fleury; Poitiers: O. Pourrat, R. Robert; Saintes: J. L. Vasquez, L. Roux; Tours: T. Boulain, A. Legras, D. Perrotin; Vierzon: F. Bandaly, D. Favreau.

Related

No related items
American Journal of Respiratory and Critical Care Medicine
157
4

Click to see any corrections or updates and to confirm this is the authentic version of record