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

Predictors of Extubation Outcome in Patients Who Have Successfully Completed a Spontaneous Breathing Trial.

Khamiees M, Raju P, DeGirolamo A, Amoateng-Adjepong Y, Manthous CA.

Chest 2001;120 1262-1270. [abstract: full-text for subscribers]

Reviewed by Joel Barron MD, UCLA, Los Angeles, CA

Review posted October 20, 2002

I. What is being studied?

Study objective:

To determine the effect of "airway competence" variables on extubation outcomes in mechanically ventilated patients who have successfully completed a spontaneous breathing trial (SBT). These variables include cough strength and magnitude of endotracheal secretions. The study also examined whether patients with a PaO2/fraction of inspired oxygen (P:F) ratio between 120 to 200 had different outcomes than patients with P:F ratio > 200 in terms of successful extubation.

Study design

The study was a prospective, unblinded cohort 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?

Ninety-one adult patients, intubated in the Intensive Care Unit (ICU) were followed until discharge from the ICU or placement of a tracheostomy. There is no mention of criteria for tracheostomy placement. The patients are only followed to the extent that they were extubated successfully or not. They were not followed necessarily to hospital discharge or death.

2. Was follow-up sufficiently long and complete?

Discharge from the ICU implies that patients were successfully extubated, whereas tracheostomy suggests that extubation was not deemed a short-term possibility. Both outcomes are fairly definitive events with regard to the study, and provide adequate data to test the hypotheses.

3. Were all potential predictors included?

The list of potential predictors considers several variables of clinical significance, and is thorough with regard to those variables relevant to the hypotheses. It is not, however, an exhaustive list.

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

Univariate analysis was performed to determine the strength of association between study parameters and extubation outcome. This analysis generated risk ratios for each predictor variable.

5. Were outcome variables clearly and objectively defined?

The main outcome variable was clearly and objectively defined as extubation success or failure. Patients who remained extubated for at least 72 hours were considered successfully liberated from mechanical ventilation. The decision to intubate and reintubate was made by the patients' clinicians; no explicit criteria or protocol was followed.

III. What are the results?

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

The prediction tools are as follows:

1. Assessment of voluntary cough strength using a somewhat observer-dependent scale of 0 to 5, and the "white card test," (WCT) a more objective measure of a patient's ability to generate a forceful cough (how well the subject was able to cough secretions onto a white card held 1-2 cm in front of their open endotracheal tube). Assessments of cough strength were converted to binary variables: weak or no cough vs. strong cough, positive or negative white card test. scant or mild secretions vs. moderate or abundant secretions, frequency of suctioning ≤ or > 2 hours. The unadjusted risk ratios for extubation failure as determined in univariate analysis are listed in the table below.

Unadjusted RRs for Extubation Failure

Variables	Number	RR	95% CI
Cough strength
Weak or no cough	24	4.0	1.8-8.9
Negative WCT result	26	3.0	1.3-6.7
Secretions
Moderate/abundant	48	8.7	2.1-35.7
Suctioning frequency ≤ 2 hrs	49	16.0	2.2-116
Hemoglobin ≤ 10 g/dl	33	5.3	2.1-13.6
Weaning parameters
P:F ratio < 200	19	0.5	0.1-2.1
RSBI ≤ 105 breaths/min/L	17	1.0	0.3-3.0

After Table 3 on p. 1266.

2. Estimate of magnitude of endotracheal secretions, based on subjective findings as well as frequency of endotracheal suctioning. Both variables correlated with extubation failure. See above for relative risk of these variables. Magnitude of endotracheal secretions were also treated as binary variables: scant or mild secretions vs. moderate or abundant secretions, frequency of suctioning ≤ or > 2 hours.

3. The synergy between these two variables, cough strength and secretions, was clearly demonstrated. Patients with weak cough and moderate/abundant secretions were 32 times more likely to fail extubation than patients with strong cough and scant secretions. The synergistic effect of the two variables was nearly four times their additive effect when relative risk was considered (Rothman synergy index 3.7, and a Hogan index 0.57).

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

In multivariate analysis, the odds ratios are large and quite convincing. Although the variables appear to be strongly associated with extubation failure, the wide confidence intervals suggest that the sample size is too small to give a very precise estimate of risk.

Variables	Adjusted Odds Ratio	95% CI
Model 1
Suctioning frequency ≤ 2 hrs	33.1	3.2 - 341.0
Weak or no cough	12.8	2.6 - 62.2
Hemoglobin ≤ 10 g/dl	7.2	1.6 - 31.7
Model 2
Moderate or abundant secretions	19.7	3.1 - 127.3
Weak or no cough	12.6	2.6 - 62.1
Hemoglobin ≤ 10 g/dl	7.4	1.8 - 30.1
Model 3
Suctioning frequency ≤ 2 hrs	24.0	2.7 - 216.0
Negative WCT result	7.3	1.7 - 31.5
Hemoglobin ≤ 10 g/dl	6.6	1.6 - 26.4

After Table 4 on p. 1266.

The table above presents three models of risk for extubation failure using the three parameters airway secretions, cough strength, and hemoglobin. The adjusted odds ratios reflect the cumulative risk based the presence of each risk factor, e.g., in Model 1 hemoglobin ≤ 10 g/dl confers an OR of 7.2, weak or no cough with hemoglobin ≤ 10 g/dl is associated with an OR of 12.8, and suctioning frequency ≤ 2 hrs, weak or no cough and hemoglobin ≤ 10 g/dl carries an OR of 33.1.

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?

The models were not tested in an independent sample of patients.

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

My patients clinically resemble the study patients in that they are at the same point in the course of disease and treatment, but are dissimilar in that they are pediatric patients. If we assume the pathophysiology of respiratory failure after extubation to be similar in adults and children, the results should be comparable.

3. Does the tool improve your clinical decisions?

The authors suggest that several factors limit the ability to generalize their results. They cite a lack of truly objective, validated measures to assess the predictive variables and an absence of strict protocols directing the decisions to extubate or reintubate (although some would suggest this improves the "real world" applicability of such a study). Additionally, the variables considered alone have neither the sensitivity nor specificity to definitively guide therapy. However, the synergy between the two variables poor/absent cough and moderate/large secretions strongly suggests that such patients represent a group at high risk for extubation failure.

The tests of secretions and strength of cough are highly predictive of success. The positive predictive value for these tests was very strong. Patients with scant or no secretions and those requiring suctioning less frequently than every two hours (noted as > 2 hours above) had an extubation success rate of 96% and 98%, respectively. The negative predictive value of these tests is only 33%, i.e., 66% of patients with lots of secretions and weak coughs will still succeed at extubation. Several studies have sought reliable indicators for favorable outcome after extubation. Where this and many other studies falter is in their inability to foresee extubation failure. One study of gastric-arterial PCO2 measurements during respiratory stress-testing showed 97% PPV and 94% specificity, but only 55% NPV. (1) In a similar study, CROP and rapid shallow breathing indices were evaluated as predictors of extubation in children. CROP index proved to be a useful tool for risk stratification, but again NPV was low at only 33%. (2) If we used these tests to delay extubation, the majority of our patients would remain on support longer than necessary.

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

The positive predictive values of the pre-extubation variables for successful extubation, 88% - 98%, are encouraging. Despite the lack of sensitivity and specificity of these tests, I believe they can provide reassurance of a favorable outcome to patients who test well.

References

1. Uusaro A, Chittock DR, Russell JA, Walley KR. Stress test and gastric-arterial PCO2 measurement improve prediction of successful extubation. Crit Care Med. 2000;28(7):2313-9. [abstract]; PedsCCM EB Journal Club Review by N Hoover, A Vats
2. Baumeister BL, el-Khatib M, Smith PG, Blumer JL. Evaluation of predictors of weaning from mechanical ventilation in pediatric patients. Pediatr Pulmonol. 1997;24:344-52. [abstract]; PedsCCM EB Journal Club Review, by S. Venkataraman

 

Comments

Back to the EB Journal Club Index