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

Evaluation of predictors of weaning from mechanical ventilation in pediatric patients.

Baumeister BL, el-Khatib M, Smith PG, Blumer JL.

Pediatric Pulmonol 1997;24:344-52. [abstract]

Reviewed by Shekhar Venkataraman MD, University of Pittsburgh, Children's Hospital of Pittsburgh

Review posted October 27, 2000

I. What is being studied?

Study objective:

Yang and Tobin, in 1991, reported that the ratio of the respiratory rate to tidal volume (f/Vt) which is an index of rapid shallow breathing (RSB) and the CROP index, which is an integrated index incorporating respiratory mechanics, a measure of oxygenation, respiratory rate, and muscle strength, were very useful in predicting weaning failure in adults (1). CROP is calculated as [Cdyn x Pimax x (PaO2/PAO2)]/RR where Cdyn is dynamic compliance, Pimax is maximal negative inspiratory pressure generated against an occluded airway, PaO2/PAO2 is the ratio of arterial pO2 to alveolar pO2, and RR is respiratory rate. RSB is calculated as f/Vt or respiratory rate divided by spontaneous tidal volume. In this study, the authors examined the usefulness of these two indices to predict extubation failure in pediatric patients.

Study design

1. A convenience sample of 47 medical and surgical PICU patients receiving mechanical ventilation and considered suitable for possible extubation were enrolled prospectively after informed consent was obtained from the parents.
2. None had a chest tube at the time of measurements and the extubation trial.
3. The PICU medical team were blinded to the results of the measurements except for arterial blood gases
4. A priori, the patients were classified into three age groups: 0-2 months, 3-36 months, and 3-22 years.
5. Using a pneumotachometer placed between the endotracheal tube and the ventilator circuit, Cdyn, RR, airway flow, and Vt were measured. In addition, the maximal negative inspiratory pressure was measured against an occluded airway.
6. Extubation for at least 24 hours was considered a success.
7. Both RSB and CROP were calculated and the tidal volume measurement was indexed to the body weight. The respiratory rate required to calculate the RSB and CROP was a total of both spontaneous and mechanical breaths.
8. Cut-off values were defined and receiver operating characteristic (ROC) curves were generated with both indices. A logistic regression analysis was performed using both indices.

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?

Probably. The cohort consisted of a mixture of ages and diagnoses; approximately half the patients had congenital heart disease though it is not clear if these were postoperative patients. All were apparently followed through the full post-extubation period (at least 24 hours).

2. Were all potential predictors included?

Not really. Only modified CROP and RSB indices were presented. Although these calculations include direct measurements of tidal volume and respiratory rate, it would have been helpful to provide this raw data as well, as others have shown that tidal volume alone (referenced to body weight) represents a reasonable predictor of extubation success or failure. (2)

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

Yes. The indices were tested both independently with ROC curves and together by logistic regression.

4. Were outcome variables clearly and objectively defined?

Mostly. Extubation success was easy to define as the need for reintubation within 24 hours of extubation. Although they commented that no patient was reintubated for evidence of airway obstruction (a common problem necessitating reintubation that would not be predictable based on the measurements made in this study), they included respiratory distress and depression as reasons for reintubation. In general, respiratory depression is likely to be caused by CNS problems or drug effects independent of respiratory mechanics, though some patients with "poor respiratory effort" may have been fatigued. There is no breakdown of the numbers of patients reintubated for respiratory distress vs. depression.

Additional comments regarding the study validity:

1. In the original description by Tobin (1) observed that the rapid shallow breathing pattern emerged almost immediately upon discontinuation of mechanical ventilator support. In the study by Yang and Tobin, CROP was also measured off all mechanical ventilator support.

2. The authors did not measure RSB and CROP according to the original description in the literature (1, 3). Both RSB and CROP were derived during complete spontaneous breathing without any ventilator support. Providing mechanical ventilator breaths during the measurement makes the derived RSB index and CROP suspect.

3. Khan's study (4) showed that in many children the fraction of the total minute ventilation provided by the ventilator is considerable, thereby masking the ability of the patient to sustain spontaneous breathing without mechanical ventilation. RSB and CROP measured off all ventilator support in this study showed very poor discriminating capacity for these two indices in children.

4. Since the original methodology was not followed, it would be very difficult to accept these data as valid. The patients in this study were on considerable amount of ventilator support when these measurements were obtained. It is critical to follow the original methodology to make this test valid. Otherwise, everyone will have their own variation of the test leading to completely variable results.

III. What are the results?

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

1. 47 patient sets were collected. There were 9 extubation failures (19% failure overall).
2. The respiratory rates observed were within 3 standard deviation from the mean value for age.
3. An RSB value of 11 bpm/ml/kg and a CROP value of 0.1 ml x mmHg/bpm/kg had the best discrimination between success and failure. All patients with a modified CROP index of greater than this value succeeded at extubation and all below this value failed. ROC curves for RSB and CROP revealed an area of 0.84 and 1.0, respectively. CROP performed much better than RSB in this study.

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

In their hands with this population, the modified CROP score was a perfect predictor of extubation success or failure. However, no confidence intervals were presented, preventing the reader from extrapolating conclusions with much certainty. In fact, with a successful extubation rate of 81% in a sample size of 47, the 95% confidence intervals for the proportion are 0.68 and 0.90 (Calculation available at: http://glass.ed.asu.edu/stats/analysis/pci.html). Using contingency table analysis with the following data is problematic for the calculations:

	Success	Fail
CROP > 0.1	38	0
CROP < 0.1	0	9

However, using the lower 95% confidence interval value for the proportion of success (0.68), the table would look like this:

	Success	Fail
CROP > 0.1	32	6
CROP < 0.1	6	3

This would provide us with the following sensitivity, specificity, positive and negative predictive values for a CROP value cut off of 0.1: 0.84, 0.33, 0.84, and 0.33. In other words, from the data provided, with 95% confidence the negative predictive value might be as poor as 33%, meaning that only 33% of patients with a CROP < 0.1 would fail extubation.

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?

This has not been studied. Therefore, it is difficult to determine the validity of these cut-off values. Without validation in an independent sample, these are still preliminary results.

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

Yes. They are.

3. Does the tool improve your clinical decisions?

No. Not at this moment.

References

1. Tobin MJ, Perez W, Guenther SM, et al. The pattern of breathing during successful and unsuccessful trials of weaning from mechanical ventilation. Am Rev Respir Dis. 1986;134(6):1111-8.[abstract]
2. Farias JA, Alia I, Esteban A, Golubicki, Olazarri FA. Weaning from mechanical ventilation in pediatric intensive care patients. Intensive Care Med 1998; 24: 1070-1075. [abstract] [PedsCCM EBJC Review]
3. Yang KL, Tobin MJ. A prospective study of indexes predicting the outcome of trials of weaning from mechanical ventilation. N Engl J Med. 1991;324(21):1445-50. [abstract]
4. Khan N, Brown A, Venkataraman ST. Predictors of extubation success and failure in mechanically ventilated infants and children. Crit Care Med. 1996;24(9):1568-79. [abstract]

 

Comments

Back to the EB Journal Club Index