Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Extubation failure in pediatric intensive care: A multiple-center study of risk factors and outcomes.

Kurachek SC, Newth CJ, Quasney MW, et al.

Crit Care Med. 2003 Nov;31(11):2657-64. [abstract]

Reviewed by Sydney Partin Primis MD; Angela T. Wratney MD MSCR; and Ira M. Cheifetz MD FAARC FCCM, Duke University Medical Center

Review posted June 18, 2005

I. What is being studied?

Study objective:

To determine a contemporary failed extubation rate, risk factors for extubation failure, and consequences of extubation failure in pediatric intensive care units. The authors state three hypotheses, none of which have been previously investigated in a multi-center, prospective study:

1. Extubation failure is, in part, disease specific.
2. Preexisting respiratory conditions predispose to extubation failure.
3. Admission acuity does not predict extubation failure.

Study design

A prospective, observational cohort study of 2,930 intubated pediatric patients in 17 diverse PICUs during a 12 month period (April, 2000 - April, 2001). Each center enrolled 18 sequential patients per month for 12 consecutive months. This allowed for seasonal variability and eliminated a potential unbalancing effect of larger units.

Inclusion Criteria: birth-18 years old; intubated ≥ 4 hours; enrolled once per hospitalization.

Exclusion Criteria: premature infants (only two centers provided care to premature babies); any patient not eligible for a planned extubation trial including patients with a tracheostomy present or planned, limitations on care (DNI/DNR), death prior to extubation, and unplanned extubation.

Outcomes analyzed

The primary outcome for this study was extubation failure as defined by the need for reintubation within 24 hrs after a planned extubation trial.

II. Are the results of the study valid?

Note: These questions follow from Randolph AG et al. Understanding articles describing clinical prediction tools. Crit Care Med 1998;26:1603-1612. [abstract]

1. Was a representative group of patients completely followed up? Was follow-up sufficiently long and complete?

Yes.

Participating PICUs: Sixteen of the 17 diverse PICUs who participated in this study completed enrollment. The mean bed number in the 16 PICUs was 21 (range: 12 - 36), and the mean number of admissions for the year prior to study initiation was 1,309 (range: 573 - 2,243).

Patient Characteristics: Due to unplanned extubations, 136 patients were excluded from the trial and were analyzed separately leaving 2,794 patients who experienced a planned extubation trial. The characteristics and subsequent analyses are limited to this patient group The study population accurately reflects the typical patient variability of most PICUs.

Available resources: The most senior in-house physicians providing nightly coverage were residents at seven sites, fellows at five sites, and attendings at four sites. Written protocols governing airway maintenance included cuffed endotracheal tube care at nine sites, ventilator weaning at two sites, an extubation procedure at five sites, and pre-extubation respiratory mechanics of any type performed on ≥ 50% of the patients at three sites.

2. Was follow-up sufficiently long and complete?

Yes. Extubation failure was defined as reinsertion of the endotracheal tube within 24 hours of a planned extubation trial. Site participants chose a post-extubation evaluation period of 24 hours as part of the study design due to general agreement that > 80% of extubation failures occur within the first 24 hours of extubation. Previous studies have examined the extubation failure rate occurring within 24, 48, or 72 hours of extubation (1-3). In this study cohort, 83% of failed planned extubation trials occurred within the first 12 hrs.

3. Were all potential predictors included?

No. Potential patient predictors associated with extubation outcome that were not tested include an assessment of: history of recent intubation, report of a difficult or anatomically narrowed airway perhaps assessed by the number of attempts required to place the endotracheal tube or the need for ancillary intubating equipment (i.e., fibroscopy), and exposure to steroids prior to the extubation trial. A subgroup analysis was not performed on those who were intubated for mechanical ventilation (i.e., ARDS) versus those intubated specifically for airway support (i.e., subglottic stenosis). In adult studies, a hemoglobin < 10 g/dl is also associated with failed extubation but has not yet been studied as a predictor in the pediatric age group (4).

As the authors indicate, they could not evaluate the impact of standardized weaning and extubation practices due to the reported absence of these policies in all but three participating PICUs. Local practices which may have affected extubation outcome include standardized weaning and sedation management during mechanical ventilation (5-7), pre-extubation respiratory mechanics measurements such as assessment of the endotracheal tube air leak (3), deadspace to tidal volume ratio (Vd/Vt) (1), or the negative inspiratory force measurement (NIFM) (8). Also due to infrequent and variable use (318 patients), the effect of non-invasive ventilation on extubation outcome could not be fully assessed in this study population.

4. Did the investigators test the independent contribution of each predictor variable?

Yes. Logistic regression was performed to assess the potential role of numerous variables. Three general patient features were found to be associated with a higher extubation failure rate: patient origin, locale of ETT placement, and the operative status of the patient. Specifically, the highest extubation failure rates were seen in those patients who were: 1) admitted from another ICU, the general floor, and from "other" locations; 2) intubated in the PICU or at an outside hospital; and 3) among pre-operative and non-surgical patients. Specific features found to be statistically associated with failed extubations included: age ≤ 24 months, dysgenetic or syndromic conditions, medical and/or surgical airway conditions, the need to replace the ETT on admission to the PICU, chronic conditions, and the need for "rescue" interventions within 24 hrs of extubation (steroids, heliox, racemic epinephrine, and non-invasive positive pressure ventilation).

5. Were outcome variables clearly and objectively defined?

Partially. Extubation failure was clearly clinically defined as the patient requiring reintubation within 24 hrs of a planned extubation trial. The decision to extubate and the need for reintubation were determined for each patient by standard local procedures. The extubation failure rate was objectively defined as the number of patients who failed a planned extubation divided by the total study population.

As part of the descriptive aspects of the study, the etiology of extubation failure was assigned to each episode of a failed extubation subjectively by the PICU attending based upon objective clinical and laboratory data. One or more of five potential causes for the failed extubation were chosen: upper airway obstruction, inadequate gas exchange, respiratory muscle weakness, hemodynamic instability, and neurologic impairment. No protocols for formal testing or objective study-defined criteria were in place to support the chosen etiology for failed extubation in an individual patient. The rate of mortality was also reported in this cohort, but it is unclear if this is death prior to PICU or hospital discharge.

III. What are the results?

1. What is(are) the prediction tool(s)?

The authors report the results of their primary outcome and a number of secondary descriptive analyses to test the following three hypotheses: A) Hypothesis driven results

1. Extubation failure is disease specific: This was true. Extubation failure occurred in 174 of 2,794 patients (6.2%; 95% CI 5.3-7.1). There was no statistically significant difference across the 16 sites (range 1.5-8.8%, p = 0.088). The observed extubation failure rate was higher for patients with acute medical (12.4%) or surgical airway (11.8%) conditions, dysgenetic (10.4%) or syndromic (14.0%) conditions, and underlying chronic conditions (10.5%).
2. Underlying respiratory conditions affect extubation outcome: This was true. The observed extubation failure rate for patients with underlying respiratory conditions was 10.5%, and for those patients requiring noninvasive positive pressure ventilation before their acute illness the rate was 26.1%.
3. Admission acuity does not affect extubation outcome. This was true. The admission PRISM II score was not statistically associated with the risk for extubation failure. The authors propose that this may be due to the scoring process being too far removed from the extubation event to accurately predict outcome.

B. Descriptive results:

1. Extubation Failure Etiology: The most frequently reported cause of extubation failure was upper airway obstruction (37.3%) followed by pulmonary dysfunction (27%), respiratory muscle weakness (13%), hemodynamic instability (9.5%), and neurologic causes (9.5%). A combination of etiologies was reportedly present in 3.6% of the patients.
2. Associated Clinical Characteristics: Of note, the following clinical characteristics were associated with a higher risk of extubation failure:
1. younger patients, particularly < 24 months of age (extubation failure rate 7.1%);
2. longer duration of intubation (148.7 ± 207.8 hours vs. 107.9 ± 171.3 hours; p < 0.001)
3. patients intubated ≥ 48 hours (8% vs. 4.2%; p < 0.001);
4. patients intubated > 10 days (extubation failure rate 17.5%);
5. patients who required immediate ETT replacement upon PICU admission (extubation failure rate 11.3%).

Notably, extubation failure was not associated with the following clinical features: race, gender, weight, underlying cardiovascular or oncologic disease, presence of acute respiratory conditions associated with the respiratory syncitial virus (apnea, bronchiolitis, or pneumonia), or the need for rescue therapies such as extracorporeal membrane oxygenation, high-frequency oscillatory ventilation, inhaled nitric oxide, or tracheal insufflation. However, due to the small numbers of patients in each group of analyses and the small number of failed extubations in this cohort, it is unclear if the lack of association amongst these variables and extubation failure is due to a Type II error or due to actual phenomena.

3. Morbidity and Mortality:

1. A prolonged PICU length of stay was observed for those patients who failed extubation vs. those who successfully extubated (17.5 ± 15.6 days vs. 7.6 ± 11.1 days; p < 0.001).
2. Twenty five patients of the failure group (n=173) (14.5%) received a tracheostomy.
3. Of those patients who initially failed extubation, 63% successfully extubated on the 2nd attempt, and 13.2% successfully extubated after the 3rd extubation attempt.
4. Seven patients died, and the risk of mortality is significantly higher in those patients who failed extubation (4%) than in those who extubated successfully (0.8%).

Of note, unplanned extubations occurred in 136 of 2,930 total patients (4.6% of the original patient population or 1.02 per 100 intubation days). The failure rate for unplanned extubation was 37.5%. Whereas 14.5% of the study patients were ≤ 6 months of age, 48.8% of unplanned extubations occurred in this population.

Five risk factors were chosen on the basis of their clinical or statistical relationship with extubation failure to be included in a multivariable analysis. Death was not chosen due to the infrequency of the event. The five factors included: (1) age (> 24 months and ≤ 24 months), (2) intubation duration (≤ 48 hrs and > 48 hrs), (3) presence of an underlying respiratory condition, (4) presence of an underlying neurologic condition, and (5) the requirement for endotracheal tube replacement upon admission to the PICU. Multivariable logistic regression revealed the odds of extubation failure associated with each predictive factor and an interaction between an underlying chronic condition and the length of ventilation:

1. age ≤ 24 months, OR 1.6 (95% CI 1.2 to 2.3)
2. need for endotracheal tube replacement upon admission to the PICU, OR 2.0 (95% CI 1.2 to 3.2)
3. presence of an underlying respiratory condition and duration of intubation ≤ 48 hrs, OR 3.1 (95% CI 1.7 to 5.8)
4. presence of an underlying respiratory condition and duration of intubation > 48 hrs, OR 1.3 (95% CI 0.8 to 2.0)
5. presence of an underlying neurologic condition and duration of intubation ≤ 48 hrs, OR 1.3 (95% CI 0.6 to 2.8)
6. presence of an underlying neurologic condition and duration of intubation > 48 hrs, OR 2.9 (95% CI 1.8 to 4.7)

2. How well does the model categorize patients into different levels of risk?

While there is no scoring system described, subsets of patients at higher risk for extubation failure are clearly identified by the odds ratios above. The model is limited by the few variables chosen to be included in the multivariable model. With 174 failed extubation events, it is unclear why the authors did not evaluate a greater number of clinically relevant predictors.

3. How confident are you in the estimates of risk?

Patients with an underlying respiratory condition and a duration of intubation ≤ 48 hrs have 1.7 to 5.8 higher odds of failing extubation than a similar patient without those risk factors. Similarly, it appears that if the same patient were intubated > 48 hrs, the odds of failing extubation are reduced to only 1.3 times average. Likewise, for the patient with an underlying neurologic condition, the data suggest these patients may have no greater odds for extubation failure than if the duration of intubation required is ≤ 48 hrs.

IV. Will the results help me in caring for my patients?

1. Does the tool maintain its prediction power in a new sample of patients?

Unknown. The authors did not test the validity of the predictive tool in a new population.

2. Are your patients similar to those patients used in deriving and validating the tool(s)?

Yes. This multiple-center study was conducted within sixteen diverse PICUs. Randolph et al (9) reported similar patient characteristics when screening patients for six consecutive months among nine large PICUs across North America. Notably, from that data Randolph reported 17.1% of patients (1,096) were intubated for a minimum of 24 hrs, and the median length of ventilation was seven days. In this trial, a remarkable number of patients were intubated > 48 hrs. In general, the duration of mechanical ventilation is relatively short, the mortality rate is very low, and the clinical characteristics of the population may be quite varied.

Other studies have also found the mortality for patients who have failed extubation is higher than for those who extubated successfully (10,11). Farias et al reported an in-hospital mortality rate of 46.4% for failed extubation as compared with 6.3% for patients extubated successfully (11). Recently, Farias et al reported results from a survey of the International Group of Mechanical Ventilation in Children among 36 PICUs. The mortality among those patients who required reintubation after a failed extubation was 19% (10).

3. Does this tool improve your clinical decisions?

Potentially. Extubation failure is associated with a higher mortality, a greater PICU length of stay, and a greater risk for future extubation failure. The results of this study help to identify a greater association of failed extubation with general patient features; however, very little information is provided which guides our interventions or identifies clinically modifiable events. One point of therapeutic interest to the clinician may be the interaction between an underlying chronic condition (such as neurologic or respiratory disease) and the length of ventilation prior to the extubation trial. A higher odds of failed extubation results when the duration of intubation is ≤ 48 hrs for patients with underlying respiratory disease and when it is > 48 hrs in patients with an underlying neurologic condition. These data may support a more conservative extubation plan for these patients or support greater vigilance in post-extubation monitoring and early supportive intervention. However, it is impossible to discern cause from effect in this analysis; in other words, was failed extubation a marker of a risk factor that may be causally related to the poorer outcomes, or was the failed extubation per se the cause? The latter seems unlikely and does not necessarily argue for keeping some patients intubated longer than clinically indicated. After all, the increased risk of failed extubation with these risk factors is very modest.

One important missing variable in this study is the role of noninvasive mechanical ventilation in supporting successful liberation from mechanical ventilation. The study is unable to capture locally successful practices, such as the use of non-invasive ventilatory support, or other interventions as a bridge to independent ventilation after extubation, early extubation after cardiac surgery, optimal use of sedatives and muscle relaxants, and standardized weaning protocols.

In this survey, the extubation failure rate for unplanned extubations was 37.5% (51/136). Stated differently, the rate of successful extubations in this group was 62.5%. This may argue, as have some authors in the past, that we are delaying extubation or prolonging the duration of mechanical ventilation unnecessarily (4,6,10-11).

4. Are the results useful for reassuring or counseling patients?

Yes. Extubation failure is a relatively rare event for our patients in the PICU. Now documented in this large multiple-center study, the extubation failure rate for our patient population can be estimated to lie between 5.3 and 7.1%. The narrow confidence interval obtained across sixteen diverse PICUs also supports the success of practices and general uniformity in outcomes with regards to the management of mechanical ventilation and extubation practices at each individual center.

Although commonly known via clinical experience, parents may be counseled that it is nearly impossible to predict the outcome of a given extubation trial for any one particular patient. The presence of the following risk factors, however, may raise the possibility that a failed extubation may occur: age < 24 months, prolonged intubation > 48 hrs and certainly if > 10 days, an underlying respiratory or neurologic condition, the need for noninvasive positive pressure ventilation prior to admission, a genetic or syndromic condition, the need for an ETT change upon PICU admission, and a medical or surgical airway condition.

The reasons for failing extubation were subjectively determined in this study; however, in 68 patients (40%) inadequate respiratory muscle strength and pulmonary dysfunction were noted. As the authors indicate, this may be amenable to monitoring, strength testing, and intervention.

References

1. Hubble CL, Gentile MA, Tripp DS, et al. Deadspace to tidal volume ration predicts successful extubation in infants and children. Critical Care Medicine 2000; 28(6):2034-2040. [abstract] [PedsCCM EBJC Review]
2. Baumeister BL, el-Khatib M, Smith PG, et al. Evaluation of predictors of weaning from mechanical ventilation in pediatric patients. Pediatr Pulmonol 1997; 24(5):344-352. [abstract] [PedsCCM EBJC Review]
3. Mhanna MJ, Yaacov BZ, Tichy CM, et al. The "air leak" test around the endotracheal tube, as a predictor of postextubation stridor, is age dependent in children. Critical Care Medicine 2002; 30:2639-2643. [abstract]
4. Epstein SK. Decision to extubate. Intensive Care Med 2002; 28:535-546. [abstract]
5. Farias JA, Alia I, Esteban A, et al. Weaning from mechanical ventilation in pediatric intensive care patients. Intensive Care Medicine 1998; 24:1070-1075. [abstract] [PedsCCM EBJC Review]
6. Ely EW, Baker AM, Dunagan DP, et al. Effect on the duration of mechanical ventilation of identifying patients capable of breathing spontaneously. New England Journal of Medicine 1996; 335:1864-1869. [abstract]
7. Randolph AG, Wypij D, Venkataraman ST, Hanson JH, et al. Effect of mechanical ventilator weaning protocols on respiratory outcomes in infants and children: A randomized controlled trial. Journal of the American Medical Association. 2002; 288:2561-2568. [abstract] [A HREF="../THERAPY/Randolph-weaning.html">PedsCCM EB Journal Club Review]
8. MacIntyre NR, Cook DJ, Guyatt GH. Evidence-based guidelines for weaning and discontinuing ventilatory support: A collective task force facilitated by the American College of Chest Physicians; the American Association for Respiratory Care; and the American College of Critical Care Medicine. Chest 2001; 120(6):375S-395S. [abstract]
9. Randolph AG, Meert KL, O'Neil ME, Hanson JH, et al. The feasibility of conducting clinical trials in infants and children with acute respiratory failure. American Journal of Respiratory and Critical Care Medicine. 2003; 167: 1334-1340. [abstract]
10. Farias JA, Frutos F, Esteban A, Retta A. What is the daily practice of mechanical ventilation in pediatric intensive care units? Intensive care medicine. 2004; 30:918-925. [abstract]
11. Farias JA, Retta A, Alia I, Olazarri F, et al. A Comparison of two methods to perform a breathing trial before extubation in pediatric intensive care patients. Intensive Care Medicine 2001; 27:1649-1654. [abstract] [PedsCCM EBJC Review]

 

Comments

Back to the EB Journal Club Index