B. P. Kavanagh
The conventional aims of mechanical ventilation are to provide adequate oxygenation, carbon dioxide (CO2) clearance and relieve work of breathing. An additional aim is the recruitment of lung tissue based on multiple animal experiments and some clinical experiments, although this could not be considered a conventional aim at present. A final aim is to prevent multiple organ failure and death, although the appropriate means or targets necessary to achieve this are not at this stage validated for use by the bedside physician. In addition to providing benefit, an aim must be to prevent harm. In this context, the clinician tries to avoid barotrauma, multiple organ dysfunction, atelectasis, hemodynamic impairment, and patient ventilator asynchrony. This chapter will review the key conventional targets and some potential future targets. The advantages of each target will be reviewed and finally the 'trade off of competing ventilation targets will be discussed. Although critically important, titration towards ventilator-patient interactions and sedation are beyond the scope of this chapter.
Titrating towards a parameter in the ICU suggests that the parameter is valid (a worthy goal), and that the clinician balances the risks and benefits of such titration. In terms of oxygenation, ICU clinicians opt for similar qualitative targets; however the quantitative variability is large . It is anticipated that among other - less 'established' - ventilatory targets, the variability may be greater.
Adequacy of tissue oxygenation is a key aim in mechanical ventilation. However, although measurement of arterial oxygen tension is a valid, reproducible standard at the bedside, the overall aim, of course is adequacy of tissue oxygenation. Thus, arterial oxygenation should be viewed as an intermediate therapeutic target 10 , that is necessary - but not sufficient - as an end-point in targeting mechanical ventilation.
Measures of arterial oxygenation can be characterized by whether they take account ofventilatoryparameters, oxygen deliveryor tissue oxygen supply balance.
Arterial oxygen saturation is continuously available at the bedside and is a sensitive and robust parameter. An oxygen content measured with a co-oximeter is more meaningful when measuring venous oxygen levels, and is currently available as a continuous parameter. Although these measures provide a good idea of the oxygen content, once the number approaches 100% there can still be a profound intrapul-monary shunt that is detectable only by changes in arterial oxygen tension. Continuous PaO2 is measurable with bedside arterial blood gas sampling, or by use of intra-arterial opthode devices. The latter technology, especially in small children or in adults with significant peripheral vascular disease, may be limited by the relative dimensions of the intra-arterial device, and the diameter of the artery. In terms of titrating, for example, inspired oxygen fraction (FiO2) or positive end-expiratory pressure (PEEP), pulse oximetry oxygen saturation (SpO2) is insensitive when above 94-96% because significant shunt can exist with PaO2 values of 9-50 kPa (I kPa = 7.5 mmHg), all ofwhich correspond to a SpO2 of 100%. Therefore, when using SpO2 for titration, the process must commence with SpO2 in the range of 90-92%, so that increases in oxygen tension can be appreciated. PaO2 can also be used, and is highly sensitive to improvements in oxygenation (calculation of the A-aO2 differences might be more sensitive to reductions in ventilation/perfusion [V/Q] mismatch, although unless automated, it is somewhat cumbersome). The limitations, in the absence of an indwelling analyzer, are the times required for sampling and laboratory (or 'point of care') processing.
The ratio of the partial pressure of arterial oxygen (PaO2) to the fraction of inspired oxygen (FiO2) was initially devised as an entry criteria definition for oxygenation impairment in ARDS by a Consensus Conference group10 . This is extremely useful in the sense that when FiO2 is altered, this alteration is factored into the picture. Thus PaO2 has far more meaning with the physician knows the value of the FiO2. In addition, although mean airway pressure is not incorporated into the PaO2/FiO2 quotient, alterations in PEEP or mean airway pressure will be reflected in a changed PaO2/FiO2 ratio. This translates into a useful index, because if the PEEP or mean airway pressure has been increased, and this increase is accompanied by a 'better', i.e., higher, PaO2/FiO2 ratio), this indicates a successful result of the pressure alteration (e.g., successful recruitment).
A potential advance on the PaO2/FiO2 ratio is the oxygenation index. This was initially developed for pediatrics, and is a composite quotient taking account of three parameters (FiO2, mean airway pressure, and PaO2), and thus provides - in theory - a more robust sense of the gas exchange efficiency of the lung. It is expressed as:
(FiO2 x mean airway pressure x 100)
Oxygenation index =--
There are several important differences between the PaO2/FiO2 ratio and the oxygenation index. First, an increased PaO2/FiO2 ratio indicates 'better' oxygenation, whereas an increase in the oxygenation index indicates a worsening of the oxygenation status. Second, where the oxygenation index is increased (worsened), this may reflect a change in either mean airway pressure or FiO2, in contrast to the PaO2/FiO2 ratio where mean airway pressure is ignored. Although the oxygenation index maybe a superior 'integrated' indexof the oxygenation efficiency ofthe lungs during mechanical ventilation, there are important gaps in our knowledge. We do not know the relative importance of incremental alterations of the three parameters, nor do we understand whether they have linear relationships. In this sense the formula is empiric, not derived. Finally, it is possible that certain conditions exist that render the oxygenation index unsuitable. For example, with excessive mean airway pressure leading to over inflation, the oxygenation index may become disproportionately higher. In such a situation, considering the oxygenation index only, without regard to its individual elements, could be interpreted as a worsening of the underlying physiology, suggesting that higher FiO2 or mean airway pressure are required, as opposed to the correct option, i.e., lowering the mean airway pressure.
Measures Incorporating Global Perfusion Global O2 delivery (DO2)
This parameter requires measurement of cardiac output and arterial oxygen content. This might be a logical target for oxygenation because it involves delivery apart from just oxygenation of blood, but it requires the presence of a pulmonary artery catheter, or other accurate measure of cardiac output. However, in many groups of critically ill patients (e.g., those with sepsis), cardiac output is extremely high. Thus, from the outset, provided SpO2 is above say 90%, it is obvious that measures to increase global DO2 do not have any bearing on normal physiology. Therefore, global DO2 is often far greater than normal values before any titration is commenced. Furthermore, the clinician is interested in local conditions at the organ or cellular level, rather than global delivery.
Mixed venous O2 saturation (SvO2)
This reflects the difference between global DO2 and global oxygen consumption (VO2) and is a surrogate marker for oxygen extraction by the tissues. While theoretically attractive as a marker of true 'adequacy' of DO2, the SvO2 suffers from several difficulties. First, any faulty assumptions - see above - that apply to DO2 apply also to SvO2. Second, there are inherent flaws due to mathematical coupling that may invalidate SvO2 as a measure of global VO2. This flaw has been demonstrated by comparing calculated vs. measured VO2 . Finally, phenomena such as flow weighted averaging (e.g., minor contributions from vulnerable vascular beds are diluted in the global cardiac output), and non-contribution of specific tissue venous outflow, can render the global SvO2 insensitive to local oxygen supply-demand imbalance.
Of all the tonometric methods, gastric tonometry has been the best validated and several studies suggest that hemodynamic management driven by gastric tonometry is alogical approach in criticallyillpatients. The principles oftonometry are that gastric mucosal intracellular pH (pHi) is reflected in the gastric muscosal intracellular CO2, which can be deduced using avariety of saline or aero-tonometric techniques. It has been suggested that correction for arterial CO2 level, such that titration against (PiCO2-PaCO2) instead of titration against PiCO2 alone, is optimal, in order to correct for fluctuations in arterial CO2 tension. Gastric tonometry would ideally represent a composite marker of metabolic demands balanced against the local supply of tissues. In principle, this would be an excellent composite measure of oxygen content, delivery and consumption at the local level. However, the caveats include adequacy of sampling, regional variation, measurement and methodologic concerns, as well as the assumption that indeed the gastric mucosa truly reflects vulnerable tissue beds. The technique has been used in weaning from mechanical ventilation, but does not replace the established markers of weaning success (e.g., f/VT ratio) . Nonetheless, the differences in intramucosal oxygen supply-demand status associated with successful vs. failed weaning attempts underline the importance of integrated systemic markers of respiratory distress .
Near Infrared Spectroscopy (NIRS)
This is a technique that measures an aggregate index of tissue oxygenation and has been standardized for a variety of tissues including the brain10 . NIRS reflects tissue oxygen status in a global manner but regional differences cannot be reliably deduced. In addition, it is not really understood what NIRS represents and the technology requires standardization on a regular basis. Although not reported in ARDS, it has been studied in neonatal respiratory failure to assess high frequency oscillatory ventilation (HFO) .
In summary, although all the above measures of oxygenation assessment are in use in critical care, there is really no good assessment as to what constitutes the optimal measure of tissue oxygenation. There does, however, appear to be a consensus that an oxygen saturation of > 90% is a reasonable target provided that it can be safely achieved. In the neonatal critical care literature, established criteria for extracorporeal membrane oxygenation (ECMO) are based on the oxygenation index, i.e., recognition ofthe concept that despite a toxic level ofFiO2 and damaging levels of mean airway pressure that the resultant PaO2 (preductal) is inadequate for the neonatal brain. In adult critical care, such criteria are not as well established. Finally, although there are multiple modalities for monitoring and titration of tissue oxygenation on the horizon, the last consensus conference concluded that clinical assessment of oxygenation relied, for now, on bedside assessment .
Was this article helpful?