INTRODUCTION

While there is relative consensus as to when mechanical ventilation should be initiated in the presence of respiratory insufficiency, the management of infants during recovery from respiratory failure remains largely subjective and is predominantly determined by institutional or individual practices or preferences. This can lead to babies either being left on the ventilator too long, or extubated too hastily, with subsequent reintubation. Therefore, a method that reliably identifies those babies ready for extubation would be a useful adjunct to the clinical decision-making process. Attempts have been made to devise some objective predictive indices that might help to identify the optimal time for extubation. However, compared to adults, there have been only a few published studies on the weaning parameters used in the neonatal population.^1,2,3 Moreover, none of these studies provide a "threshold value" for individual measurements that could consistently discriminate between success and failure of extubation.

We have used a relatively new method, the "minute ventilation test" (MVT), to assess a baby's readiness for extubation. This test measures the effectiveness of spontaneous breathing and respiratory muscle endurance, and can be obtained from real-time pulmonary monitoring, now routinely available on most modern ventilators. In a previous observational study, carried out in two separate units, we found that this method had a positive predictive value of 86%.^4,5 This prompted us to test this method further in a prospective, randomized, controlled manner.

METHODS

Infants were eligible for enrollment in the study if they had a diagnosis of respiratory distress syndrome requiring surfactant replacement therapy and assisted mechanical ventilation for at least 24 hours. The study was performed in a level III neonatal intensive care unit. Randomization to enter the study was carried out when the infant had reached the "weaning stage," defined by predetermined criteria of FiO₂ <0.4, peak inspiratory pressure (PIP) <16 cmH₂O and mean airway pressure (P_AW) <10 cmH₂O. Parental consent was obtained in each case and the study was approved by the institutional research ethics committee.

Assignment of the method of extubation was performed with block randomization using sealed envelopes. Babies were assigned to either extubation by conventional methods based on the clinical discretion of the attending clinician, or by use of the MVT. For the latter, the mechanical minute ventilation was averaged over a period of 10 minutes while the baby was still receiving full ventilatory support from flow-cycled, pressure-limited, assist-control (MV_M). Minute ventilation was attained using a pneumotach placed proximally to the endotracheal tube, the information from which was fed to an on-line pulmonary graphics computer. This calculated the minute ventilation as a product of the expiratory tidal volume and respiratory rate as measured by the pneumotach. The ventilator was then changed to CPAP (range 3 to 4 cmH₂O) with no (control) breath rate, and the minute ventilation during spontaneous breathing (MV_S) was averaged over the next 10 minutes. The baby was extubated if MV_S/MV_M was 50% and no adverse effects (such as apnea, bradycardia, or an increased oxygen requirement) were encountered during the period on CPAP. If the baby failed the initial MVT, it was performed again 6 to 8 hours later and repeated at the same interval until the baby achieved the required criterion. All the babies in the study were ventilated using VIP BIRD^® Infant/Pediatric Ventilator (Bird Products, Palm Springs, CA), and were generally prescribed caffeine (20 mg/kg loading dose, followed by 5 mg/kg every 24 hours) prior to testing where appropriate.

The primary outcome measure for both treatment and control groups was the time from randomization until "successful" extubation, defined as remaining extubated for more than 24 hours. Data were computed and statistical analysis performed using the Mann–Whitney U-test and Kaplan–Meier survival plot with log-rank analysis.

RESULTS

A total of 42 babies were enrolled in the study. They were equally distributed between the two groups and had similar demographic characteristics (Table 1). There was a statistically significant reduction in time from randomization to extubation in those babies evaluated by the MVT compared to clinical assessment only (mean time of 8 hours versus 36 hours; p=0.04) (Table 2). Although five out of 21 babies required reintubation within 24 hours in the MVT group compared to two out of 21 in the control group, this difference was not significant (Table 3). Reasons for reintubation are given in Table 4.

Table 1 - Demographic Details of the Two Groups.

Full table

Table 2 - Time from Randomization to Extubation.

Full table

Table 3 - Ventilatory Status at 24 Hours Postextubation.

Full table

Table 4 - Reason for Reintubation.

Full table

Among the whole study population, including both treatment and control groups, infants who needed reintubation were more immature (mean gestational age 27 weeks versus 30 weeks; p=0.036) and tended to weigh less at birth (mean weight 1.0 kg versus 1.6 kg; p=0.059). Four babies in each group were not prescribed caffeine, with one infant in the MVT group requiring reintubation. Five infants received postnatal dexamethasone (three in the MVT group and two in the clinical group).

DISCUSSION

Although an astute clinician might be able to predict the time that an infant is ready to start weaning, extubation failure still occurs in about a third of cases.⁶ Therefore, any attempt to devise some objective predictive indices that could help to identify the optimal time for extubation and lead to reduction in the duration of invasive ventilation or the need for excessive reintubation would be desirable towards interrupting the pulmonary injury sequence. Therapeutic adjuncts, such as caffeine⁷ and dexamethasone^8,9 have been used to facilitate this process, but because of safety concerns, they are, at present subjects for further investigation.

The MVT has been used to assess a baby's readiness for extubation. This method is advantageous because of its ease of application by using real-time data already provided by the ventilator, and because of the ability to perform this test serially in a noninvasive way apparently without jeopardizing a baby's safety.

In this study, the positive predictive value of the MVT for extubation was 76% (95% CI, 55 to 89%), which is very close to the value noted in the original observational study. It was not practical to blind the present study, and the longer duration of ventilation from time of entry into the study until extubation in the clinical decision (control) group may have resulted from clinician bias. There is also a possibility that the observed differences in outcome between the two groups may have been magnified by the small sample size.

There were seven babies who failed extubation in our study. Two resulted from apnea, both of which were in the MVT group. If apnea is obstructive in origin, the MVT will fail to detect this while the baby is still ventilated with an endotracheal tube. The other five babies were reintubated for a worsening biochemical picture from respiratory acidosis. It was noted that there seemed to be a trend toward a lower MV_S /MV_M ratio in those babies who did not achieve successful extubation. Maintenance of minute ventilation in a range sufficient to assist adequate removal of carbon dioxide is a prerequisite to successful extubation, and alveolar hypoventilation and subsequent respiratory acidosis will occur if either the tidal volume or respiratory rate is too low. It is possible that the discriminatory value of the MV_S/MV_M ratio in this study was small. There may also be several other factors responsible for extubation failure, which a single test of pulmonary mechanics such as the MVT may fail to recognize, for example, postextubation laryngeal edema, apnea, or thick secretions.

The MVT reliably predicted those babies who were ready for extubation earlier than clinical judgement alone. The findings of this study could be used either to refine this method in order to improve its discriminatory power or combine it with other objective measures, which would help the clinician to accurately assess a baby's capability for independent ventilation prior to extubation.

References

1. Dimitriou G, Greenough A, Endo A, Cherian S, Rafferty G. Prediction of extubation failure in preterm infants. Arch Dis Child Fetal Neonatal Ed 2002;86:F32–5.
2. Balsan M, Jones J, Watchko J, Guthrie R. Measurements of pulmonary mechanics prior to the elective extubation of neonates. Pediatr Pulmonol 1990;9:238–243.
3. Kavvadia V, Greenough A, Dimitriou G. Prediction of extubation failure in preterm infants. Eur J Pediatr 2000;159:227–231. | PubMed |
4. Wilson BJ Jr., Donn S, Sinha A, Sinha S. Does spontaneous minute ventilation predict readiness for extubation in mechanically ventilated preterm infants [Abstract]. Pediatr Res 1998;43:310A.
5. Wilson BJ Jr., Becker M, Linton M, Donn S. Spontaneous minute ventilation predicts readiness for extubation in mechanically ventilated preterm infants. J Perinatol 1998;18:436–439.
6. Sinha S, Donn S. Weaning from assisted ventilation: Art or science? Arch Dis Child Fetal Neonatal Ed 2000;83:F64–70.
7. Henderson-Smart D, Steer P. Methylxanthine treatment for apnea in preterm infants. [update of Cochrane Database Syst Rev.2000;(2):CD000140; 10796304.] [Review][17 refs]. Cochrane Database Systematic Rev 2001;3:CD000140.
8. Halliday H, Ehrenkranz R. Early postnatal (<96 hours) corticosteroids for preventing chronic lung disease in preterm infants. [update of Cochrane Database Syst Rev.2000;(2):CD001146; 10796252.]. [Review] [59 refs]. Cochrane Database Systematic Rev 2001;1:CD001146.
9. Halliday H, Ehrenkranz R. Delayed (>3 weeks) postnatal corticosteroids for chronic lung disease in preterm infants. [update of Chronic Database Syst Rev.2000;(2):CD001145; 10796251.]. [Review] [35 refs]. Cochrane Database Systematic Rev 2001;2:CD001145.