Introduction

The acute respiratory distress syndrome (ARDS) is characterized by an alveolar blood-gas barrier injury. Conventional mechanical ventilation may have detrimental effects in patients with severe ARDS [1]. The recent ARDS Network study has proven the efficacy of a protective low tidal volume ventilation strategy [2]. However, this ventilation concept may induce severe hypercapnia. Hypercapnia can be deleterious in some patients and is contraindicated in cases of metabolic acidosis and severe brain injury [3, 4]. In some situations it is necessary to protect the blood-gas barrier and concomitantly to limit the hypercapnia level. Two main solutions have been proposed: tracheal gas insufflation and/or increased respiratory rate [5, 6, 7, 8]. Another simple solution may be the instrumental dead-space reduction, as already reported by Hurni et al. [9]. This simple method is frequently mentioned, though it has not been thoroughly evaluated in adult ARDS patients.

We designed a prospective open study to evaluate the efficacy of a humidification system dead-space (DSh) reduction in order to decrease PaCO2 levels in adult ARDS patients.

Material and methods

Patients

Patients were included in the study if they were intubated and mechanically ventilated using a low tidal volume (mean tidal volume below 8 ml/kg of ideal body weight), and already had an indwelling radial artery catheter. ARDS was defined according to recently published guidelines: sudden onset of respiratory failure, bilateral diffuse infiltrates on the chest X-ray and arterial oxygen tension/fractional inspired ratio (PaO2/FIO2) below 200 mmHg [10]. Exclusion criteria were: chronic respiratory insufficiency and/or chronic CO2 retention and all situations in which PaCO2 levels may vary independently of the dead-space reduction protocol (hemodynamic instability and/or major bronchial secretion). The Simplified Acute Physiologic Score (SAPS II) [11] and the Logistic Organ Dysfunction score (LOD) [12] were calculated 24 h after ICU admission. The Lung Injury Severity score (LIS) was calculated within 24 h of ARDS onset [13]. The study was approved by the Ethics Committee of the Société de Réanimation de Langue Française as requiring no informed consent.

Protocol

Sedation was maintained by a continuous infusion of titrated midazolam and fentanyl, in order to obtain a Ramsay's sedation score value of 6 [14]. All patients were ventilated in the volume control mode with a constant tidal volume and ventilatory settings remained constant during the measurement periods (NPB 840, Tyco healthcare, USA). The study consisted of five periods of 30 min that were sequentially tested in each patient. The following five conditions were sequentially tested in each patient: (1) heat and moisture exchanger (internal volume=95 ml) with a tracheal closed-suction system (internal volume=25 ml, i.e. a total humidification system dead-space volume [DSh=120 ml], the usual humidification system policy in our ICU), (2) heat and moisture exchanger (internal volume=45 ml) with the closed-suction system (DSh=70 ml), (3) heat and moisture exchanger (internal volume=25 ml) with the closed-suction system (DSh 50 ml), (4) heated humidifier with the closed-suction system (DSh=25 ml) and (5) heated humidifier alone (DSh=0 ml) (see Fig. 1).

Fig. 1.
figure 1

Representation of the five phases of the protocol. (DSh humidification system dead space, HME heat and moisture exchanger, HH heated humidifier, CSS tracheal closed-suction system)

Full size image

Tracheal suction was performed after each humidification system modification. Subsequently, no further changes in ventilatory parameters, vasoactive or inotropic drug infusion variations, or fluid administration were allowed during the study periods. Mean systemic arterial pressure, cardiac rhythm and pulse oximetry (SpO2) were continuously monitored (SC 9000, Siemens, Sweden). Patients were to be withdrawn from analysis if SpO2 became equal or inferior to 90% despite FIO2 increase, if mean arterial pressure became equal or inferior to 60 mmHg or in case of any additional tracheal suctioning requirement, fluid administration or vasoactive and/or inotropic drug infusion. Arterial blood gases, hemodynamic and ventilatory parameter measurements were always performed after a 30-min stability period, consecutively to each humidification system dead-space reduction.

Measurements

Arterial blood gas samples were always performed using the indwelling radial artery catheter. All blood gas measurements were performed using an IL 1620 monitor (Chiron diagnostic) and the results were corrected for the patient's core temperature. Five measurement sets were performed for each patient. Ventilatory parameters were recorded using the NPB 840 ventilator monitoring system: tidal volume (VT), respiratory rate (RR), extrinsic positive end-expiratory pressure (PEEPe), intrinsic positive end-expiratory pressure (PEEPi) and plateau pressure (Pplat). Hemodynamic parameters were recorded during each study period: heart rate (HR) and mean arterial pressure (MAP).

Statistical analysis

An independent statistician carried out the statistical analysis. Data are expressed as means ± standard deviation and were analyzed using a one-way analysis of variance for repeated measures (ANOVA). Significance between the five study periods was determined by Fisher's protected least significance test. A p value equal to or less than 0.05 was considered statistically significant.

Results

Patients characteristics and baseline measurements

Eleven patients were included in the study. One patient was excluded from analysis because of hemodynamic instability, according to the previously determined withdrawal criteria. The characteristics of the remaining ten patients are summarized in Table 1. Eight patients were receiving a continuous infusion of catecholamines for associated septic shock. Two patients were receiving inhaled nitric oxide. The overall survival rate was 30%.

Table 1. Patients' characteristics. Results expressed as means ± SD
Full size table

Basal measurements for respiratory parameters are summarized in Table 2. Mean tidal volume (VT=6.9±1.8 ml/kg; extreme values: 4.7–9.5 ml/kg) and mean initial respiratory rate (RR=20±1 breaths/min; extreme values: 16–21 breaths/min) remained constant throughout the entire measurement periods.

Table 2. Ventilatory, gasometric and hemodynamic parameters measured at inclusion in the study. Results expressed as means ± SD
Full size table

Effects of the humidification system dead-space reduction

A significant and progressive pH increase was observed throughout the artificial dead-space reduction procedure: from 7.18±0.08 at DSh120 to 7.28±0.08 at DSh0 (minus 120 ml artificial dead space; p<0.05). A significant PaCO2 decrease, from 80.3±20 mmHg at DSh120 to 63.6±13 mmHg at DSh0, was also observed (p<0.05). A statistical difference was observed at DSh50, DSh25 and DSh0, compared to the initial DSh120 values, for both pH and PaCO2 (p<0.05) (Table 3). Otherwise, no variation was observed for PaO2/FIO2, Pplat, intrinsic PEEP or any hemodynamic parameter during any of the five study periods.

Table 3. Ventilatory, gasometric and hemodynamic parameters measured during the five phases of the study. Results expressed as means ± SD
Full size table

Discussion

This study demonstrates that PaCO2 levels, resulting from a protective ventilation strategy in ARDS patients, can be maintained within acceptable limits using a very simple instrumental dead-space reduction maneuver, using a heated humidifier rather than a heat and moisture exchanger.

The efficacy of the instrumental dead-space reduction on PaCO2 levels is due to an alveolar minute ventilation increase, as already demonstrated by Campbell et al. [15]. In intubated and mechanically ventilated patients, instrumental dead space is mainly composed of the endotracheal tube, the heat and moisture exchanger and the Y-piece connector. In our study, the maximal instrumental dead-space volume reduction was obtained by simply switching from our usual heat and moisture exchanger device with the concomitant use of a tracheal closed-suction system, to a heated humidifier without closed-suction system (minus 120 ml). The study protocol did not include a tracheal tube length reduction, as its internal volume is too small (2 cm of a 8 mm internal diameter endotracheal tube represents a 3 ml internal volume). A slight PaCO2 decrease was observed when switching from the heated humidifier combined with a closed-suction system to the heated humidifier alone. Since the PaCO2 decrease associated with closed-suction system removal was minimum, and in order to use the system with the best efficacy/ergonomic ratio, we decided to keep the closed-suction system in routine practice, to limit alveolar deflation related to endotracheal suctioning [16].

Compared to tracheal gas insufflation (TGI)-induced PaCO2 decrease, such a 20% reduction, when simply reducing the instrumental dead space, seems to be less significant. In hypercapnic dogs with oleic-acid-induced lung injury, Nahum et al. observed that expiratory washout with 10 l/min flow oxygen allowed a 29±5% PaCO2 decrease [7]. The same results were observed by various authors in ARDS or studies involving patients with lung disorders when using similar procedures [5, 8]. Whereas a 27% PaCO2 decrease was observed with TGI (15 l/min) by Richecoeur et al. in six ARDS patients (initial PaCO2 84±24 mmHg), a respiratory rate increase (from 18 to 30 breaths/min) achieved a similar PaCO2 reduction, and combined methods allowedverall 46% PaCO2 reduction when applied concomitantly [6]. Prin et al., in a similar study (instrumental dead-space reduction 100 ml), observed a 18% PaCO2 decrease [17]. In our study, the PaCO2 decrease is directly proportional to the instrumental dead-space volume reduction. Combining the results of all these studies, it thus appears that each of three methods allows a significant PaCO2 reduction, and these three methods may be applied in combination. The main differences between the methods are their clinical feasibility or their potential adverse effects.

Tracheal gas insufflation decreases PaCO2 but has several limits: (1) the PaCO2 decrease is directly proportional to the oxygen flow rate (it requires an oxygen flow of 6–15 l/min) [7], (2) difficulties with the optimal catheter position maintenance, in order to obtain the maximal PaCO2 decrease (1 cm above the tracheal bifurcation; maximal PaCO2 decrease obtained in Nahum et al. study, compared to 5 or 10 cm above the carina) [7], (3) mean airway pressure increase, due to an expiratory flow limitation [5], (4) difficulties in obtaining air-tightness, (5) the necessity of using a specific device (expiratory washout flow generator), which must be synchronized with the expiration time [6].

Increasing respiratory rate results in a minute ventilation increase, which thus explains the PaCO2 decrease. However, the optimal respiratory rate, i.e. the respiratory rate that allows a significant PaCO2 decrease without increasing intrinsic PEEP and the risk of barotrauma, may be difficult to assess. While increasing the respiratory rate in 14 patients with acute lung injury, Vieillard-Baron et al. did not observe any PaCO2 decrease, whereas an intrinsic PEEP increase and a significant drop in the cardiac index occurred [18].

Many authors have studied the effects of humidification devices on respiratory parameters in spontaneously breathing patients. In those studies, using heat and moisture exchangers instead of heated humidifiers has already been shown to induce ventilatory and arterial blood gas parameter variations. Those variations were explained by inspiratory and expiratory resistance increase (responsible for a work-of-breathing increase), by an alveolar ventilation decrease and/or by intrinsic PEEP increases [19, 20, 21, 22]. These observations must be taken into account when analyzing the present study results, even if we did not observe any intrinsic PEEP variation in our patients, who were sedated, paralyzed and ventilated in a control-mode.

A non-randomized study with fixed sequence of conditions has, however, some limits, of which the main ones are the effects of time and the effects of sequence. Moreover, we can not exclude that a systematic tracheal suctioning every 30 min after each humidification system modification had some influence on the results. Finally, all those questions may also be typical French or European problems because, whereas heat and moisture exchangers are routinely used in our ICUs, heated humidifiers alone are preferentially used worldwide [23].

In conclusion, artificial airway dead-space reduction allows a significant PaCO2 reduction. Independently of any respiratory mechanical changes, this very simple maneuver may be of importance when low tidal volume ventilation and PaCO2 lowering are warranted.