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

Specific questions in this review are based on Jaeschke RZ, Meade MO, Guyatt GH, Keenan SP, Cook DJ. How to use diagnostic test articles in the intensive care unit: diagnosing weanability using f/Vt. Crit Care Med. 1997;25(9):1514-2. [abstract]

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Deadspace to tidal volume ratio predicts successful extubation in infants and children.

Hubble CL, Gentile MSA, Tripp DS, Craig DM, Meliones JN, Cheifetz IM.

Crit Care Med 2000; 28:2034-2040. [abstract]

Reviewed by Al Torres, MD, MS, University of Illinois College of Medicine at Peoria

Review posted December 15, 2000

I. What is being studied?

Study objective:

To identify a minimal physiologic deadspace (Vd/Vt) value using single-breath carbon dioxide (CO2) capnography for predicting successful extubation from mechanical ventilation in pediatric patients.

Study design:

A prospective, blinded, non-interventional clinical study. 30 minutes before planned extubation, each subject was ventilated with pressure support ventilation to deliver an exhaled tidal volume of 6 ml/kg. After 20 minutes of pressure support ventilation, an arterial blood gas value was measured, Vd/Vt was obtained from the CO2SMO Plus respiratory profile monitor (computed by the monitor by averaging Vd/Vt for a minute's worth of breaths), and the subjects were extubated. Over the next 48 hours, the clinical team managed the subjects without knowledge of the preextubation Vd/Vt values.

II. Are the results of the study valid?

Primary questions:

1. Was there an independent, blind comparison with a reference standard?

Yes. The reference standard for a Vd/Vt of < 0.50 being predictive of extubation success was absence of the postextubation clinical outcome of respiratory distress requiring reintubation or noninvasive ventilation. Respiratory distress included the following: increased work of breathing (e.g., nasal flaring, accessory muscle retractions, grunting, or increased respiratory rate for age), hypoxia (identified by pusle oximeter or ABG), or hypoventilation (confirmed by ABG). Noninvasive ventilation included one of the following modes: a) nasal CPAP; b) nasopharyngeal CPAP; c) mask BiPAP; and d) negative pressure ventilation delivered with a chest wall curass.

The use of subjective criteria reflect current practice in weaning PICU patients making these criteria an adequate reference standard. The clinical team was blinded to the Vd/Vt value of each subject for a minimum of 48 hours to reduce bias in the assessment of respiratory distress. We have no way of knowing if there was any bias in the people deciding who needed post-extubation support and who didn't. Use of more objective criteria to decide who should receive postextubation support would have eliminated this potential bias.

2. Did the patient sample include an appropriate spectrum of patients to whom the diagnostic test will be applied in clinical practice?

Maybe. 45 children, ranging in age from 1 week to 18 years (mean age 43 +/- 62 mo), were enrolled. Gender was almost evenly split (24 females, 21 males). 47% (21/45) of the subjects had primary lung disease such as pneumonia ( n = 12), ARDS (n = 2), sepsis (n = 1), asthma (n = 1), pulmonary hemorrhage (n = 1), pulmonary hemosiderosis (n = 1), ingestion (n = 1), acute chest syndrome (n = 1), and near drowning (n = 1). 18 additional patients underwent cardiothoracic surgery (17 for cardiac repair, 1 for tracheoesophageal fistula repair). The remaining 6 patients were either trauma patients (n = 3) or postoperative from spinal fusion surgery (n = 3).

Although the investigators did study a wide disease spectrum, they failed to provide information confirming (or refuting) a range of lung severity (e.g., PF ratio, oxygenation index) or overall illness severity (e.g., PRISM, PIM) in the population being tested. If the authors only selected patients with severe disease, this lack of severity spectrum may result in a falsely high sensitivity or specificity.

Secondary questions:

3. Did the results of the test being evaluated influence the decision to perform the reference standard?

No. All patients who underwent Vd/Vt testing were extubated by the clinical team without knowledge of the results. Verification bias exists if the result of the diagnostic test (Vd/Vt) influences whether study patients undergo confirmation with the reference standard (clinical assessment for signs of respiratory distress postextubation)(1,2).

III. What are the results?

1. Are the test's sensitivity, specificity, and likelihood ratios presented (or are the data necessary for their calculation provided)?

Yes. The investigators selected a Vd/Vt of 0.50 as the cutoff point. The sensitivity and specificity of the Vd/Vt compared with the extubation outcome were 0.75 and 0.92, respectively (75% of patients who successfully extubated had a Vd/Vt < 0.50 and 92% of patients who failed extubation had a Vd/Vt > 0.50). The positive predictive value and negative predictive value were 0.96 and 0.60, respectively (96% of patients with Vd/Vt < 0.50 successfully extubated, and 60% of patients with a Vd/Vt > 0.50 failed extubation).

The likelihood ratio of a Vd/Vt < 0.50 in a patient with a successful extubation (LR for a positive test is true positive rate/false positive rate or sensitivity/1 - specificity) was 9.8 (95% CI, 1.5 to 64.8). The likelihood ratio a Vd/Vt < 0.50 in a patient with extubation failure was 0.27 (95% CI, 0.15 to 0.50). In other words, patients who succeed at extubation are 9.8 times as likely to have had a Vd/Vt < 0.5 than patients who fails extubation, or, alternatively, patients who fail extubation are 0.27 times as likely to have a Vd/Vt < 0.5 as those who succeed at extubation.

The investigators further analyzed the 20 subjects with Vd/Vt > 0.5 into 2 subgroups around a second cutoff point, 0.65.

Vd/Vt	successful extubation	extubation failure	likelihood ratio
< 0.5	24	1	9.8
0.5 - 0.65	6	4	0.6
> 0.65	2	8	0.1

The LR does not allow you to predict directly the chance of an outcome occurring - it must be used in conjunction with the pre-test probability by a nomogram or a formula. So in this case, their overall need for respiratory support was 29% after extubation - the pretest probability of failure. Then using the LR of 0.27 and the nomogram on p 1517 of the Jaeschke article (3), we see the post-test probability would be about 15%. So with a Vd/Vt of < 0.5, the probability of extubation failure is approximately 15%. If we change the definition of extubation failure to include patients requiring reintubation only, the pretest probability is 6.7% and the post-test probability of extubation failure is 2%. The trouble with this study and so many like it is that the test is very good at predicting success but not as good at predicting failure, especially when a bad outcome is infrequent.

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

1. Were the methods for performing the test described in sufficient detail to permit replication?

Almost. Although the investigators were quite specific about the subjects' respiratory mechanics and ventilator settings at the time of the Vd/Vt measurements, the investigators do not mention how many minutes of Vd/Vt were averaged to obtain the single value used or whether the value used was the value that occurred most often over a specific period of time (i.e., steady state achieved). The use of of pressure support ventilation with a tidal volume of 6 ml/kg may affect the validity of the Vd/Vt measurements. For example, an increase an alveolar ventilation secondary to recruitment may decrease ventilation-perfusion mismatch and decrease Vd/Vt and vice versa.

2. Will the reproducibility of the test result and its interpretation be satifactory in my setting?

Unknown. The investigators do not state test-retest reliability of the instrument (i.e., does the investigator get the same value when he repeats the test 5 min later?). It is likely that if the respiratory monitor was not malfunctioning, was not subject to measurement drift during the measurement time period, or imposed a change in the subject's respiratory mechanics (i.e., increased resistance) over time, the measurement is likely reproducible.

3. Are the results applicable to my patients?

Maybe. The wide spectrum of disease types makes the patients similar to those found in a multidisciplinary PICU. However, the investigators gave no indicator of the spectrum of illness severity.

4. Will the results change my management?

Maybe. The decision of whether to wean and extubate patients who by other criteria (e.g., clinical examination, PF ratio, oxygenation index, Vt indexed to body weight for a spontaneous breath) are indeterminate whether they would successfully extubate, a Vd/Vt < 0.50 or > 0.65 might influence my decision to attempt extubation. It would be helpful to know how much the posttest probability would change by adding Vd/Vt measurements in PICUs already using a combination of criteria to assess extubation readiness.

5. Will patients be better off as a result of the test?

Maybe. An aggressive approach to weaning and extubating PICU patients that included measuring Vd/Vt early and often as part of the evaluation process might result in less time ventilated. There does not appear to be any risk or toxicity associated with Vd/Vt measurements. It does require an arterial blood gas for calculation which may increase the prevalence of arterial catheters or arterial punctures in a given PICU.

References

1. Begg CB, Greenes RA. Assessment of diagnostic tests when disease verification is subject to selection bias. Biometrics 1983; 39:207-215. [abstract]
2. Gray R, Begg CB, Greenes RA. Construction of receiver operating characteristic curves when disease verification is subject to selection bias. Med Decis Making 1984; 4:151-164. [abstract]
3. Jaeschke RZ, Meade MO, Guyatt GH, Keenan SP, Cook DJ. How to use diagnostic test articles in the intensive care unit: diagnosing weanability using f/Vt. Crit Care Med. 1997;25(9):1514-2. [abstract]

 

Comments

Back to the EB Journal Club Index