Mechanical ventilation has traditionally been utilized to assist in the appearance of carbon dioxide as a respiratory waste gas. Limitation ofVT or inspiratory pressure to protect the lungs against excessive mechanical stretch was introduced, initially by Wung et al.  for neonatal hypoxic respiratory failure, and subsequently by Hickling et al. [10, 11]. In these situations, the PaCO2 was allowed to become elevated, and the idea that mild to moderate hypercapnia was not necessarily a harmful entity but was worth tolerating in order to spare physical ventilator-induced lung injury (VILI) became popular. This was termed 'permissive hypercapnia' . The limitations to hypercapnia have been reviewed in detail [12, 13] and the technique of permissive hypercapnia does appear to be associated with improved patient outcome, although this is not proven. The ranges of permissive hypercapnia that should be tolerated are difficult to assess. In Hickling's studies PaCO2 values of over 13 kPa were noted in some patients (with concomitant pH values below 7.1) [10,11]. In the ARDSNet study where low stretch ventilation was utilized, the mean CO2 was higher in the low stretch group than in the high stretch group . However, although statistically significant, the magnitude of the between-group difference was small, making the effect of PaCO2 difficult to interpret.
A new concept has arisen in critical care investigation termed 'therapeutic hypercapnia'. This is defined as the deliberate induction ofhypercapnia with thepotential for therapeutic benefits over and above those that might be associated with a reduction of lung stretch [15, 16]. Multiple mechanisms have been suggested whereby elevated PaCO2 per se might be associated with benefit in the critically ill.
Currently, laboratory - not clinical - evidence only exists suggesting that therapeutic hypercapnia might be beneficial. This however, must be seen in the light of potential detrimental effects. In terms of permissive hypercapnia, the primary target is Vt or airway pressure limitation, and the CO2 is not seen as a target as much as a measure to be tolerated. Several clinical studies have shown beneficial effects of ventilatory strategies that involve the development of mild to moderate arterial hypercapnia [10, 11, 14, 17, 18]. Conversely, traditional studies of ARDS associated with limitation of VT or inspiratory pressure have not shown a beneficial effect of hypercapnia, and one study suggested (although without direct evidence) that hypercapnic acidosis may increase the incidence of renal failure .
The experimental evidence relating to therapeutic hypercapnia can be divided into supportive and nonsupportive evidence. The supportive evidence is as follows:
• Pulmonary: a variety of lung models demonstrate that ambient hypercapnic acidosis protects against reperfusion injury , stretch induced injury in vitro, stretch induced injury ex vivo  and in vivo  as well as reperfusion injury in vivo .
• Central nervous system: In terms of brain protection, there is evidence that additional CO2 protects against experimental neonatal cerebral ischemia .
• Cardiac: Nomura et al. have demonstrated that hypercapnic acidosis is associated with improved myocardial performance following ischemia-reperfu-sion .
In addition, the use of pH-stat CO2 management during cardiopulmonary bypass (associated with increased administration of CO2) results in improved cardiac and neurological parameters in children undergoing cardiopulmonary bypass for correction of congenital heart disease .
These potentially positive effects must be balanced by the emerging evidence of adverse effects. First, Holmes et al. have demonstrated that although hypercapnia increases retinal oxygenation, it may also increase neovascularization ofthe retina and thus predispose neonates to retinopathy of prematurity . Although not the primary focus of their study, the more striking finding was the far higher mortality associated with the administration of CO2. While this might not occur in adult subjects, or be relevant during mechanical ventilation, it is extremely disturbing and is thus far unexplained. Second, the issue of nitration is important. Nitration, resulting from the reaction of peroxynitrite with proteins (characteristically, on tyrosine residues), sometimes results in an alteration of protein function. Zhu et al. have demonstrated that in cultured epithelial cells, introduction of carbon dioxide increases nitration of surfactant protein-A and results in a worsening of the ability of that protein to aggregate lipids . Important caveats include the fact that the solutions were all buffered to normal pH, and the study was not so much one of hypercapnia, as opposed to correction of hypocapnia. In addition, Lang et al. have demonstrated similar findings - increased nitration - in cultured epithelial cells stimulated by lipopolysaccharide (LPS) . These last two studies may be particularly important because of the data demonstrating that nitrotyrosine formation in patients with ARDS is common, and is associated with adverse effects on complex plasma proteins involved in free radical scavenging and thrombosis . Thus, while titrated hyper-capnia may be tolerated by clinicians, it may not be universally tolerated by their patients; in the context of ARDS or critical illness these caveats must be borne in mind.
Hypercapnic acidosis is accompanied immediately by a degree of tissue buffering with chloride shift and elevation of extracellular bicarbonate. Following this, renal correction takes place and the pH is buffered toward 7.4. However, metabolic acidosis and/or renal failure are common in the critically ill and buffering mechanisms are frequently non-functional or inefficient. The situation can be worsened with dilutional acidosis from resuscitation with bicarbonate free fluids or infusion of total perenteral nutrition. Nonetheless, buffering has been recommended by many authors but its clinical value is unproven. In fact, there is evidence that it may be harmful. First, buffering of hypercapnic acidosis in ex vivo lung ischemia perfusion is associated with worsening of injury, i.e., ablation of the beneficial effects of hypercapnic acidosis in this model . Second, although perhaps not applicable to hypercapnic acidosis, buffering of metabolic acidosis is associated with augmentation of introduction of endogenous acids  as has been demonstrated experimentally  and in diabetic ketoacidosis in humans. Finally, in vitro effects demonstrate that CO2 augmented protein nitration may have been enhanced by buffering of the experimental medium [29,30].
In summary, there are advantages and disadvantages associated with experimental models of hypercapnia. Whereas moderate levels of hypercapnia have been associated with improved outcome, the hypercapnia is as a result of primary reductions of tidal pressure or volume. The potential for adverse outcome, and the lack of clinical study, suggests that whereas permissive hypercapnia may be an appropriate strategy, clinical application of deliberate (therapeutic) hypercapnia is not currently appropriate.
Whereas hypercapnia may have direct or indirect beneficial effects as indicated above, hypocapnia has very few beneficial effects except in the setting of life threatening incipient brain stem herniation or (in neonates, particularly) critical pulmonary hypertension. Aside from these two definite beneficial applications , there are multiple adverse effects including tissue ischemia, impaired oxygenation and oxygen delivery, increased metabolic demand, bronchospasm, surfactant inactivation . Thus, whereas hypercapnia may or may not be a worthwhile tradeoff against mechanical injury, hypocapnia in the absence of a specific indication is never an appropriate clinical target.
Was this article helpful?