Special Feature
What do we Know About Permissive Hypercapnia?
By Charles G. Durbin, Jr., MD, FCCM
The approach to mechanical ventilation in patients with the acute respiratory distress syndrome (ARDS) has changed over the past several years. Recognition of the potential harm that may be caused by high tidal volumes has resulted in providing smaller tidal breaths, usually between 5 and 8 mL per kilogram body weight. This "lung protective ventilation" is often accompanied with enough PEEP to prevent alveolar collapse between breaths (i.e., PEEP above the lower inflection point identified during a slow pressure volume curve), limiting peak pulmonary distending pressure to less than 35 or 40 cm H2O, and allowing adequate time for exhalation to control auto-PEEP. A number of small studies in several age groups have suggested that this approach improves outcome in terms of survival and the duration of mechanical ventilation, although, some studies have not shown benefits (see Table). Recently, the lung protective approach was shown in a large multicenter study to result in reduced mortality in patients with severe ARDS.1
| Table-Some Human Studies Evaluating Outcome from Lung Protective Ventilation Which Included Permissive Hypercapnia |
||||||
| Author | Year | Number of Subjects | Average Max PaCO2 | PaCO2 Range | Effect on Outcome | |
| Hickling HG, et al. Intensive Care Med 1990; 16:372-377 |
1990 | 50 | 62 | 40-129 | Survival better than predicted by APACHE II scores and historical controls |
|
| Hickling HG, et al. Crit Care Med 1994; 22:1568-1578. |
1994 | 53 | 66.5 | 38-158 | Survival better than predicted by APACHE II scores |
|
| Amato MB, et al. N Engl J Med 1998; 338:347-354. |
1998 | 53 | 58.2 ± 3 | NA | Shorter mechanical ventilation, better 28-day outcome, same survival to discharge |
|
| Stewart TE, et al. 1998 N Engl J Med 1998; 338:355-361. |
120 | 54.4 | 28-116 | NA | Slightly worse survival than controls |
|
| Tibby SM, et al. Intensive Care Med 1999; 158:42-45. |
1999 | 28 (Peds) | 53.2 | 40-82.6 | No difference in survival from controls |
|
| Mariani G, et al. Pediatrics 1999; 104:1082-1088. |
1999 | 49 (Neo) | NA | 45-55 | Premature infants treated with surfactant, shorter ventilation period |
|
| ARDS Network. N Engl J Med 2000; 342:1301-1308. |
2000 | 861 | 44 ± 12 | NA | 25% less mortality in protective lung ventilation group, not specifically designed to include hypercapnia |
|
A consequence of using lung protective ventilation may be an abnormally high PaCO2 and, consequently, a low arterial blood pH. The term "permissive hypercapnia" has been used recently to describe this elevated PaCO2 which is the direct consequence of reduced minute ventilation, usually resulting from lung protective ventilation in ARDS patients. Hypercapnia, however, has been studied in other clinical situations over the past 50 years and has been accepted as "harmless" by most investigators evaluating lung protective ventilation. In this paper I will review what we know about permissive hypercapnia and its development and safety. To preface this discussion, most of our understanding is inferential, based on experience with hypercapnia occurring as a byproduct of another clinical intervention rather than used alone as a primary therapeutic approach. This is understandable as most clinicians view "normal" physiology as the desired endpoint of treatments. Most of the safety information about hypercapnia has been gleaned from failures of clinical interventions to maintain normality in CO2 excretion.
My first experience with hypercapnia was pivotal and personal. In 1977, while in fellowship training, I watched a 1 kg premature baby (27 weeks gestation) born without surfactant, struggle to breathe in an isolate flushed with 100% oxygen. The parents were adamant in their opposition to tracheal intubation and mechanical ventilation but hoped that the child would survive and ultimately thrive under her own power. This was at a time when survival in this size group was estimated to be about 25% and the complications of intubation and mechanical ventilation were known to be significant and severe. The infant remained severely tachypneic for several days with sternal and chest retractions and her PaO2 remained about 40 torr. High FiO2 and routine treatments were provided including breast milk by gavage. By the fifth postnatal day, her PaCO2 had risen to 75 torr with an appropriately low pH. The next day it was 90 torr. At this point, a family conference was held and the parents agreed to allow intubation and manual (hand) ventilation for a period of 24 hours to give the baby a rest if the CO2 rose any higher. Hand ventilation was recommended to reduce the likelihood of high airway pressures from a mechanical ventilator and the injury that it might induce. The next PaCO2 was 73 torr and intubation was deferred.
Over the next 10 days, the PaCO2 remained above normal, ranging from 55-75 torr, and respiratory function progressively improved. The remainder of this child’s hospital course was uneventful and she was discharged to home three months after birth having doubled her initial weight. She had sustained no cerebral hemorrhages and had no discernable lung disfunction. Her eye exam demonstrated moderate retinopathy of prematurity and she was developing normally otherwise. This child had survived a period of severe hypercarbia lasting several weeks and had experienced essentially no complications from it. Specifically, she had no evidence of intracranial pathology.
I said this experience was personal as well as pivotal. This is the true story of my oldest daughter. She graduated from college cum laude last year and is pursuing a career in anthropology. In this anecdotal report, severe, sustained hypercarbia did not result in any identifiable direct injury. Most of the data that has supported the use of "permissive" hypercapnia comes from anecdotal reports rather than systematic study in humans.
Hypercapnia and Severe Asthma
Severe asthma is the most common disease that has led to many of the reports of patients managed with permissive hypercapnia. What is most remarkable about these reports is that survival occurred and there were few, if any, clinically detectable complications from hypercarbia per se. When bronchospasm is severe enough, achievement of a normal PaCO2 is impossible. Overt barotrauma, primarily pneumothorax, can be life threatening in these patients. Clinicians have suggested intubation, sedation and paralysis, using ventilation only to maintain oxygenation but not normal PaCO2. The term "permissive hypercapnia" was coined in treating these patients. Survivors with PaCO2s greater than 160 torr have been reported with no significant complications. With this approach to severe asthma, mortality from the disease may be reduced.2,3 The limits and consequences of extreme hypercapnia have been studied in the past in patients under general anesthesia.
Apneic Oxygenation and Hypercapnia
Studies in the 1940s and 1950s demonstrated that if the body and blood stores of nitrogen are removed by breathing 100% oxygen for 30-40 minutes, hypoxia can be prevented during prolonged apnea. The effects of 100% oxygen breathing and apnea were investigated to support the practice of creating a still surgical field during general anesthesia for delicate airway surgery. Currently, this data supports the concept of preoxygenation (or rather, denitrogenation) prior to induction of anesthesia and administration of a neuromuscular blocker. Should establishment of an airway prove to be difficult, preoxygenation is believed to provide a period of time, perhaps up to 20 minutes, before the development of lethal hypoxia. The PaCO2, however, will rise substantially in these apneic patients.
Studies of apneic oxygenation in humans were carried out following a prolonged period of oxygen breathing and lasted for up to an hour.4 During the apneic period, oxygen was permitted to flow by the airway and some degree of entrainment was possible. PaCO2 values in the range of 130-250 torr were reported in these healthy patients during this technique. Hypercarbia was associated with a moderate increase in blood pressure in all patients and an occasional extrasystole in some. All patients recovered uneventfully, although they required a period of hyperventilation to remove tissue stores of carbon dioxide. Earlier, Comroe and Dripps reported two patients who were maintained apneic for up to three hours using tracheal insufflation of oxygen following denitrogenation.5 One of these patients achieved a PaCO2 of 314 torr by the end of apnea with no morbid effects. These early studies of extreme conditions are offered as support for the relative safety of permissive hypercapnia to less profound levels. These patients were healthy by today’s standards. No attempts were used to correct the concomitant acidosis.
Current anesthesia practice of providing the patient with several minutes of oxygen breathing prior to induction of general anesthesia and neuromuscular blockade probably provides only several additional minutes of safety prior to the development of significant hypoxia should ventilation prove impossible. To gain the hour or more of apnea with satisfactory oxygenation, at least 30 minutes of oxygen breathing is needed. The rate of rise of PaCO2 during apnea was determined to be about 2-3 torr per minute following the initial rise of 4-6 torr in the first minute, due to equilibration between arterial and venous levels. This is a useful fact to know. I confirmed this fact early in my practice.
I was "moonlighting" as an attending in the university ER and a patient arrived who had collapsed, was unresponsive and apneic. She was elderly, had a history of a previous stroke, was hypertensive, and had fixed and dilated pupils. We quickly intubated her, manually ventilated her with oxygen and her pupils initially shrank but soon became fully dilated again, indicating brain herniation despite hyperventilation—a uniformly fatal condition. The first-year internal medicine resident wanted to rush her to CT scan, but I felt it was hopeless as she remained apneic and unresponsive. We decided to place her on a t-piece and let nature take its course. After 45 minutes, during which time she did not breathe, she remained hemodynamically stable with a normal heart rate and blood pressure. The resident was concerned that she was not brain dead and must have been breathing, even though we could not observe it. I suggested we obtain a blood gas to confirm apnea. We calculated that with 45 minutes of apnea the PaCO2 would be: 20 (due to hyperventilation) + 45 minutes x 2.5 torr/min, or about 132 torr. The lab called back STAT to report her PaCO2 was 126 torr, a critical value! Because we had placed her on a t-piece with oxygen flowing, she continued to entrain oxygen and her PaO2 was 330 torr. We interrupted the "apneic oxygenation," placed her on room air, and over the next several minutes she became bradycardic and then asystolic, probably due to hypoxia (pulse oximeters had not yet been invented). This case confirms the predictable rise in PaCO2 with apnea as well as demonstrates the minimal impact of significant hypercapnia on vital signs.
Tracheal Insufflation of Oxygen
Transtracheal jet ventilation has been suggested as a method for preventing death should manual ventilation be ineffective and tracheal intubation fail. Tracheal insufflation of oxygen could be used in such an airway emergency as it may prevent hypoxia, provide some ventilation, and requires less complicated equipment than jet ventilation. In dogs breathing only with room air prior to apnea, low flow oxygen through catheters placed near the carina resulted in a rise in PaO2 to 200-300 torr during the first 15 to 20 minutes and a reduction in the rate of rise of PaCO2 during apnea.6 PaCO2 rose to 200 to 300 torr with complete recovery if mechanical ventilation was instituted at this time (1-2 hours). Dogs that were allowed to continue to death survived an average of 4.5 hours and died from (presumed) cardiovascular collapse with a PaCO2 greater than 400 torr and a PaO2 greater than 100 torr. Cardiovascular stability was maintained until near death. Tracheal oxygen insufflation may be useful in the emergency situation provided there is a route for the gas to exit the lung.
Transtracheal gas insufflation (TGI) during mechanical ventilation has been suggested as a technique to reduce PaCO2 during protective ventilation, by providing tracheal dead space washout and lung risk "free" ventilation. Intermittent, exhalation gas flow would seem to be safer as any flow during inhalation would add to tidal volume and increase airway pressure. Systematic examination of the factors affecting efficiency of TGI in limiting hypercapnia during treatment of ARDS has recently been published.7,8
Permissive Hypercapnia and
Intracrania Pathology
While the above discussions suggest that hypercapnia is innocuous in most patients, the patient with increased intracranial pressure (ICP) or neurologic injury may not tolerate the increased cerebral blood flow (CBF) that results from hypercapnia. If ICP is already dangerously elevated or intracranial compliance diminished, a rise in CBF may cause intra- or extracranial brain herniation from a further increase in ICP. Some patients may develop or worsen preexisting cerebral edema with the obligatory rise in CBF which accompanies a rise in PaCO2. At least five patients have been reported developing worsening cerebral edema, and one experiencing a non-aneurysmal subarachnoid hemorrhage temporally related to institution of permissive hypercapnia.9 Monitoring of jugular venous oxygen saturation (SjvO2) may allow identification of excess CBF in patients at risk of these complications. When SjvO2 exceeds 55-75%, hypercapnia should be reduced if cerebral edema or increased ICP becomes an issue.10 Hypercapnia was thought to antagonize the effects of neuromuscular blockers but this does not appear to be correct, clinically. The duration of action and magnitude of relaxation for both depolarizing and non-depolarizing relaxants is not altered, however, resting muscle tone may be increased with severe hypercarbia.
Intrinsic (Potential) Benefits of Hypercapnia
So far in this review, hypercapnia has been addressed as a potential problem and issues of the safety of increased PaCO2 have been identified. The fact is that improved patient outcome has occurred when using hypercapnia, although it has been used in combination with other potentially beneficial therapies. This does not rule out the possibility that hypercapnia may confer independent benefits of its own.11
We know that hypercapnia increases cardiac output, primarily by reducing systemic vascular resistance. This may improve vital organ perfusion and help prevent organ failure. Acidosis shifts the hemoglobin dissociation curve to the right, facilitating unloading of oxygen to the tissues. Acidosis induced by the elevated PaCO2 freely affects intracellular pH unlike metabolic acidosis, which is primarily an extracellular factor. This intracellular acidosis may result in decreased contractility of cardiac muscle, reducing oxygen demand, thus protecting a vulnerable heart from ischemia. In animal models of ischemia, myocardial blood flow is improved in all heart regions by hypercarbic acidosis and there is no evidence of an ischemic steal syndrome.12 The brain may also benefit from an elevated PaCO2 directly as well as from the augmented blood flow, if ICP elevation is not an issue limiting its use. Tissues made more acidotic by increased PCO2 produce fewer toxic oxygen free radicals as compared to tissues made acidotic with metabolic acids or those without acidosis. This could mean less reperfusion injury following ischemia.
Acidosis attenuates many inflammatory processes. With uncompensated hypercapnia there is reduced leukocyte superoxide formation, decreased phospholipase-A activity, reduced neuronal apoptosis, and less expression of cellular adhesion molecules.13 These effects could reduce organ injury and prevent organ failure. Acidosis also reduces calcium ion responsiveness, possibly contributing to myocardial cell protection. Hypercarbia inhibits xanthine oxidase, which may improve lung function by reducing oxidative injury.14
The implications of these and other observations remain speculative, although many deleterious effects are known to occur with the hypocapnia that occurs from hyperventilation. The safety observed with profound hypercapnia may result from avoiding the negative effect of hypocarbia, or it may prove to be because of the independent benefits conveyed. Much research remains to be performed before we will know the true impact of permissive or perhaps "therapeutic" hypercapnia.
References
1. ARDS Network. N Engl J Med 2000;342:1301-1308.
2. Darioli R, Perret C. Am Rev Resp Dis 1984;139:
385-387.
3. Feihl F, Perret C. Am J Resp Crit Care Med 1994;150: 1722-1737.
4. Frumin MJ, et al. Anesthesiology 1959;20:789-798.
5. Comroe JH Jr, Dripps RD. JAMA 1946;130:381-383.
6. Slutsky AS, et al. Anesthesiology 1985;63:278-286.
7. Imanaka H, et al. Am J Resp Crit Care Med 1999;159: 49-54.
8. Kalfon P, et al. Anesthesiology 1997;87:6-17.
9. Rodrigo C, Rodrigo G. Am J Emerg Med 1999;17:
697-699.
10. Tasker RC, Peters MJ. Intensive Care Med 1998;24: 616-619.
11. Hickling KG, Joyce C. Acta Anaesthesiologica Scandinavica. Supplementum 1995;107:201-208.
12. Arellano A, et al. Crit Care Med 1999;27:2729-2734.
13. Laffey JG, Kavanagh BP. Lancet 1999;354:1283-1286.
14. Shibata K, et al. Am J Resp Crit Care Med 1998;158: 1578-1584.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content