During mechanical ventilation, the application of positive pressure with a peak inspiratory pressure (PIP) in excess of 50–60 cmH₂O may result in a barotrauma [1]. Gross overinflation leads to rupture of the airways which may result in pneumothorax, pneumomediastinum or subcutaneous emphysema.

Animal experiments have established that even PIP at a level of 30–40 cmH₂O results in lung overinflation and may cause pulmonary interstitial edema, inflammation and elevated vascular permeability, a picture that resembles acute respiratory distress syndrome (ARDS) or acute lung injury (ALI) [2,3,4,5]. The pulmonary injury is often distributed quite heterogeneously. This means that normally functioning parts of the lung are scattered between parts that are diseased, either as totally consolidated lung or as collapsed potentially expandable lung [6].

Conventional ventilation may lead to overdistention of the normally functioning lung while expanding collapsed parts. Thus, mechanical ventilation may exacerbate the pulmonary pathology and/or delay recovery. In two studies of patients with ARDS, it was concluded that survival is better when high ventilation pressures are avoided [7,8]. There have been no such studies in patients with acute respiratory failure without ARDS.

Pressure-regulated volume control (PRVC) is a new mode of ventilation that combines the advantages of the decelerating inspiratory flow pattern of a pressure-control mode with the ease of use of a volume-control (VC) mode.

The aim of this study was to test the hypothesis that PRVC associated with improvements in respiratory mechanics, outcome and length of intensive care unit stay when compared with conventional VC ventilation.

The study was approved by the ethics committee of Copenhagen (reg no 01-182/93) and was conducted in accordance with the Helsinki declaration. Data were collected prospectively, and randomization was achieved using scaled envelopes.

The inclusion criteria were acute respiratory failure, a partial pressure of arterial oxygen (PaO₂; mmHg)/FiO₂ ratio of < 300 with a PEEP of 5 cmH₂O, FiO₂ between 0.4 and 0.6, and age 18 years or more. Patients with an expected ventilator therapy of less than 24 h were excluded as were patients with intracranial pathology.

After an initial stabilization period (up to 8 h) the patients were randomized to PRVC or VC. The following variables were instituted for both groups during ventilation: tidal volume (VT) 5–8 ml/kg, a respiratory rate to achieve the desired partial pressure of arterial carbon dioxide (PaCO₂), inspiratory time 30% (no pause) and constant PEEP. After a period of 2 h, the patients were switched to the alternative method of ventilation (PRVC or VC) for a further 2 h without any other ventilatory changes. After measurement, the patients were returned to the mode of ventilation initially assigned. The patients assigned to PRVC were weaned on volume support and the patients assigned to VC were weaned on pressure support. To ensure that the two groups of patients received the same kind of ventilator management a protocol was used (which can be obtained from the authors on request).

All patients were ventilated using a Siemens Servo 300 ventilator (Siemens Elema AB, Solna, Sweden). The following variables were measured: PIP, mean airway pressure, PaO₂, PaCO₂, pH and mean arterial pressure (MAP). The number of days on the ventilator, failure of the assigned mode of ventilation (PIP > 50 cmH₂O) and survival were also recorded.

Airway pressures were recorded on the display monitor of the ventilator and arterial blood gases were measured using an arterial blood gas analyser (ABL 2; Radiometer, Copenhagen, Denmark).

Forty-Four patients with acute respiratory failure were included. No patient was excluded after randomization.

There were no statistically significant differences between demographic data for the two groups: median age 57 years (95% confidence interval 52–66) and 60 years (95% confidence interval 55–70), an APACHE II score of 18 (95% confidence interval 16–22) and 16 (95% confidence interval 14–19) and a male/female ratio of 16/6 and 10/12 for the PRVC and VC groups, respectively.

The results of the crossover study are shown in Table 1, where the number of patients in each group is given as 44 because every patient was crossed over (paired comparison). There was a significantly lower PIP in the PRVC group (P < 0.0001).

This study shows the advantage of using the PRVC mode for ventilation during acute respiratory failure. PIP was lower for all patients using the PRVC mode compared to the VC mode, and alveolar ventilation was unchanged as indicated by the constant PaCO₂. The new mode of ventilation did not improve outcome or duration of treatment, despite a statistically significant difference in peak pressures (4 cmH₂O). Though this difference in peak pressure is small, it may be more relevant in situations where larger tidal volumes are contemplated.

This study is the only study that has measured the difference between PIP on the two modes of ventilation, PRVC and VC. In contrast to other studies comparing pressure-limited ventilation with various forms of ventilation, our patients had only mild respiratory insufficiency. In the studies by Rappaport et al [9] and Hickling et al [7], the inclusion criterion was a PaO₂/FiO₂ ratio < 150, whereas < 300 was used in our study. However, two of our patients failed the assigned VC mode, while no patient on PRVC failed (P = 0.24). For this tendency to have achieved statistical significance, 110 patients should have been enrolled in the study. The risk of type 2 error for an overlooked difference of 10% failing VC is only 20% with 1-β = 80%.

Further studies are needed to decide if PRVC improves outcome when compared with VC in patients with acute respiratory failure. Other subgroups, such as acute severe asthma or ARDS, could be a focus for attention.

A recent review recommends pressure-control ventilation in all clinical circumstances requiring artificial ventilation [10]. During PRVC, as with pressure control, there is a maximum pressure difference between the ventilator and the lung at the beginning of the inspiratory cycle. The resulting flow is also maximal. With the increase in intrathoracic pressure this difference diminishes, as does the resulting inspiratory flow. The flow pattern is therefore called decelerating inspiratory flow. In VC ventilation, there is a constant inspiratory flow and the resulting intrathoracic pressure is always increasing. Pressure-regulated ventilation is therefore capable of delivering the same volume at a lower PIP. This fact may play a more significant role when higher tidal volumes are required, and greater differences in peak pressures between PRVC and VC may be expected.

Our conclusion is that, during mechanical ventilation for acute respiratory failure, PIP was significantly lower on PRVC than VC, and thus PRVC may be superior to VC in certain patients.

Our results emphasize one of the basic problems in intensive care research-that therapeutic signals are too weak to be discovered in clinical trials consisting of few patients.