In this issue of Intensive Care Medicine, Farias and colleagues [1] present the results of a multicenter prospective cohort study describing children mechanically ventilated for 12 h or more in one of 36 pediatric intensive care units (PICUs) over a 2-month period from April to May 1999. This study [1], performed on behalf of the International Group of Mechanical Ventilation in Children (IGMVC), is important because it is unclear if reports of data from surveys asking PICU physicians how they think they practice represent reality.

Some of the important findings of the IGMVC study [1] were that, although synchronized intermittent mandatory ventilation (SIMV) with or without the addition of pressure-support ventilation (PSV) was the most commonly used mode of ventilation (39%), 25% of patients received pressure-control ventilation (PCV) and 23% received volume-limited ventilation-assist control mode. This is the first prospective, multicenter study to describe modes of ventilation that were being used prior to weaning. Caregiver justification for using certain modes was not described. Although high-frequency ventilation was rarely used (1.5%), it would be interesting to know what modes of ventilation were used specifically in patients with acute lung injury and acute respiratory distress syndrome (ARDS). The IGMVC study [1] also corroborated the findings of multiple other studies that the overall extubation failure rate in children is between 15 and 20% [2, 3, 4, 5, 6] and that those patients who fail and require reintubation have a higher mortality rate.

There are three important limitations to the IGMVC study [1] that the authors touch on. First, 30/36 (83.3%) of the PICUs were from Spain, South America or Central America. It is unclear how well the patient population and clinical practice mirrors that of PICUs in other developed nations. Second, the authors collected data over 2 months and missed the viral lower respiratory infection season. This would have markedly influenced their patient population. Third, the study was performed in 1999, prior to the publication of the ARDS Network study [7] that supported use of low tidal volumes and pressure limits in adult ARDS patients. The management of ARDS patients in the IGMVC study showed that, although the PaO2/FiO2 ratios were low (median values <125 on days 1–6), the amount of PEEP applied was also relatively low (median 4–8 cmH2O on days 1–6), the median tidal volumes were similar to those used in the high tidal volume arm of the ARDS Network study [7] (approximately 11 ml/kg) and the peak inspiratory pressures were somewhat high (≥40 cmH2O in 25% of patients). It is unclear whether current practice in 2004 has markedly changed from that in 1999 after publication of the ARDS Network trial results showing that the use of such high tidal volumes could lead to a worse outcome [7]. In addition, although a fairly high proportion (31%) of children in this study [1] underwent spontaneous breathing trials, the rate may be markedly higher now that more data have been published that support the benefit of spontaneous breathing trials in adults [8, 9] and children [2, 4, 10].

The Pediatric Acute Lung Injury and Sepsis Investigators’ (PALISI) Network [11] performed a similar, large, prospective cohort study over 6 months of children requiring MV for 24 h or more across nine North American PICUs. There are some important findings from this IGMVC study [1] that corroborate the findings of the PALISI study [11]. The IGMVC study found that 35% of PICU admissions required MV for 12 h or more and the PALISI study found that 1,096/6,403 (17%) of admissions required MV for over 24 h. It seems clear that the great majority of children in PICUs do not require MV over 24 h. Both studies showed that the median total duration of MV in the population of children in PICUs on the ventilator for over 12 h (IGMVC) and over 24 h (PALISI) was 6–7 days. Both studies also found that the median age of children requiring MV was approximately 1 year. Both studies found that MV modes were variable and that the majority of children were exposed to more than one mode of ventilation. Both studies showed that approximately 15% of children required MV for pneumonia and 14% for neurologic reasons. Both studies showed that the diagnosis of ARDS was uncommon (2% in IGMVC and 7.6% in PALISI) but that a high proportion of patients without the diagnosis of ARDS had PaO2/FiO2 ratios of 200 or less for a prolonged period of time.

There are also important differences between the findings of the IGMVC study [1] and the PALISI study [11]. The overall mortality rate in the IGMVC study was 15% but was only 1.6% in the PALISI study. The mortality rate for ARDS patients was 50% in the IGMVC study and only 4.3% in the PALISI study. The major difference between the two studies was that the PALISI study excluded patients (10% of total) who had life-support restrictions upon study entry and also excluded other disorders with very poor prognoses. The PALISI study was performed over the viral respiratory season and bronchiolitis was the most common single diagnosis; mortality in bronchiolitis patients is very low. The PALISI study probably best represents the population of patients who would be eligible for clinical trials comparing MV strategies, whereas the IGMVC study represents the entire cohort of patients on the ventilator.

To understand current clinical practice and how it changes over time, we need more studies like the study by Farias and colleagues [1]. Describing “standard care” and justifying MV strategies is extremely important, as failure to do so temporarily halted the important work of the ARDS Network investigators [12]. Although we do not know with accuracy how children are mechanically ventilated in most PICUs in 2004, the recent studies performed by the IGMVC [1] and the PALISI network [11] have helped to describe the population of children who require MV in PICUs. This information will assist in the design of robust clinical trials to determine eventually how we should optimally be mechanically ventilating critically ill children.