© 2004 American Thoracic Society
Therapeutic Hypercapnic AcidosisPushing the EnvelopeUniversity of Washington Seattle, Washington Should our experience and acceptance of permissive hypercapnia with lung-protective ventilation lead us to intentionally impose hypercapnia (therapeutic hypercapnia) in patients with acute lung injury? In this issue of the Journal (pp. 46–56), Laffey and colleagues point us in that direction (1). They added 5% CO2 to the inspired gas of rabbits with endotoxin-mediated lung injury to generate PaCO2 values of 70–75 mm Hg and pH values as low as 7.10. Not only was this well tolerated, but a myriad of physiological, biochemical, and histological measures of lung damage and inflammation were considerably attenuated, and mortality was reduced. Although others have shown that inspired carbon dioxide (12–25%) given prophylactically reduces ischemia–reperfusion (2, 3) and ventilator-induced lung injury (4, 5), this study (1) is the first to show a benefit when hypercapnia is administered after injury is established. Until recently, it was standard practice to correct hypercapnic acidosis with high volumes and pressures, if necessary, owing to perceived dangers of respiratory acidosis itself, especially hemodynamic depression. Evidence for harm has always been weak, as Laffey and coworkers (1) nicely illustrate. Their control and sick hypercapnic rabbits had higher blood pressures, lesser lactic acidosis, and better oxygenation than did the normocapnic controls. This remarkable tolerance to acute hypercapnia was initially noted in the 1950s when the first blood gas electrodes revealed surprisingly profound hypercapnia (PaCO2 above 150 mm Hg, pH less than 7.00) in successful thoracic surgical cases (6). A PaCO2 of 375 mm Hg secondary to prolonged inadvertent hypoventilation with complete recovery has been reported (6). Permissive hypercapnia—purposefully limiting tidal volumes and accepting hypercapnia—did not gain credence until studies in status asthmaticus were undertaken in the 1980s (7). Death rates of 20% in historial controls were reduced to almost zero largely by preventing pneumothoraces and hemodynamic depression, even with PaCO2 values above 100 mm Hg. Acute hypercapnic acidosis is well tolerated as long as perfusion and arterial oxygenation are maintained. This is due to active intracellular pH defense mechanisms that pump hydrogen ion out into the extracellular space and a neuroendocrine sympathetic response that maintains cardiac output and blood pressure despite direct negative inotropic and vasodilating effects of hypercapnia (6, 8). Although acute hypercapnia is tolerable and permits use of low tidal volumes and pressures, the concept that hypercapnia might have direct therapeutic effects in injury states, as reported by Laffey and coworkers (1), arose from studies in the 1970s and 1980s revealing a ‘pH paradox’. In tissue culture and isolated organ studies, including the lung (reviewed in [8, 9]), recovery after anoxia and/or ischemia was always greater if resuscitation occurred in an acidic milieu (pH < 7.00) and, conversely, was worse with alkalotic resuscitation. These studies established that metabolic and respiratory acidoses blunt many oxidative and inflammatory cascades initiated when tissues are reoxygenated. Because proteins generally have pH optima near physiologic pH, it is not surprising that acidosis reduces oxygen and nitrogen radical species generation, diminishes proinflammatory cytokine production, impairs neutrophil chemotaxis, and inhibits many proteases, nucleases, and phospholipases activated in injured cells (8, 9). Acting on these data, Shibata and coworkers (2) imposed prophylactic hypercapnia with 12–25% inspired CO2 begun before the onset of damage in ex vivo perfused rabbit lungs undergoing ischemia–reperfusion injury or oxidant-induced injury. They showed that features of microvascular permeability were mitigated with acidosis. Other investigators in ex vivo and in vivo studies subsequently demonstrated favorable compliance, gas exchange, cytokine generation, oxygen radical formation, and histological results (reviewed in [9, 10]). Laffey and colleagues (1) applied this strategy at a lower dose (with a lung-protective tidal volume of 4.5 ml/kg) in a clinically relevant model of bacterial-related lung injury (intratracheal endotoxin administration). They found that 5% inspired CO2, started either before (prophylactic) or 30 minutes after intratracheal endotoxin administration (therapeutic), was equally effective. They report considerably better physiological parameters (improved gas exchange, compliance, and airways resistance), more normal histology (decreased inflammatory cell infiltration and alveolar/tissue edema), and favorable biochemical results (reduced tissue and epithelial lining fluid nitric oxide metabolite concentrations, and only a slight increase in nitrotyrosine formation) with inspired CO2 even after the onset of injury. The favorable nitric oxide data are welcome because others using more intense acidotic regimens (hydrochloric acid infusions or 15% CO2) have warned of increased nitric oxide production and tissue injury (11). Lastly, the impressive results of Laffey and coworkers (1) were associated with a significant reduction in mortality (33 versus 11%). The fact that 5% inspired CO2 is efficacious when started in the midst of inflammation is an encouraging, clinically relevant finding because most often we are able to intervene only after the onset of lung damage. Yet, therapeutic hypercapnia, whether by hypoventilation or low concentrations of inspired CO2 (as administered by Laffey and coworkers [1] in healthy animals), may not be an easy strategy to initiate or sustain in critically ill patients, many of whom have underlying chronic illnesses and acute organ dysfunction that may limit their tolerance and ability to compensate against hypercapnia. Paramount is the fact that only sterile, noninfectious injuries have been studied, and only over several hours in otherwise healthy animals. Given that hypercapnic acidosis results in a broad-based partial suppression of many events critical to microbial killing, will it be beneficial in septic states? Might it not increase the likelihood of secondary hospital-acquired infections? It is mandatory to investigate this question with live bacterial challenges, both blood and airway borne. Animal models that simulate patients with compromised cardiac, renal, and cerebral function should be studied to determine whether such individuals can mount and tolerate the resulting strong neurosympathetic tone and known cardiovascular responses evoked by hypercapnic acidosis, which may cause myocardial oxygen supply–demand insufficiency, renal vasoconstriction, antidiuresis, cerebral vasodilation, and pulmonary hypertension (9, 10)? Should hypercapnia prove therapeutic, should we administer it by hypoventilation or by addition of inspired CO2 as did Laffey and coworkers (1)? Inspired CO2 might be the better choice. First, Fiehl and coworkers (12) showed that permissive hypoventilation (6 versus 10 ml/kg) leads to a greater shunt fraction and impaired gas exchange. Second, because the lung is not homogenously injured, the few remaining, more compliant units may still be over-ventilated (high ventilation–perfusion ratio) and have lower regional PACO2 values (less than 20 mm Hg) and higher local, possibly alkalotic and injurious (13), tissue pH, despite a lower minute ventilation. If the intention is to keep lung uniformly acidotic to suppress local inflammatory processes, then inspired CO2 guarantees that all regions, at a minimum, will never have a PCO2 below the inspired value. These questions of both clinical efficacy and best implementation strategy will require well planned and careful prospective studies, but given the provocative preclinical data of Laffey and coworkers (1), the answers are worth pursuing. FOOTNOTES Conflict of Interest statement: E.R.S. has no declared conflict of interest. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
