Background - Methods - Results - Characteristics of Included Studies - References - Data Tables & Graphs
Following a report by Avery et al (Avery 1987) which highlighted the variance in rates of BPD among neonatal centres, attention was refocussed upon differences in methods of treating acute respiratory disorders in the preterm infant. They postulated that the risk of BPD may be lowered if barotrauma is reduced by the early use of nasal continuous positive airway pressure (CPAP), and by accepting an arterial carbon dioxide level (PaCO2) above the physiologic range. This was the first suggestion that the incidence of BPD might be associated with low PaCO2 levels during the acute management of neonatal respiratory disease. To test this hypothesis, a multicentre historical-cohort analysis of 235 infants with birth weights of 751 to 1000 grams was undertaken by Kraybill et al (Kraybill 1989). In this study population, a PaCO2 of less than 40 mm Hg at 48 or 96 hours of life was the best predictor of BPD. Furthermore, among the 10 participating neonatal centres, the mean PaCO2 in those infants receiving mechanical ventilation at 48 and 96 hours strongly correlated with BPD rates. Garland et al (Garland 1995) performed a retrospective cohort study to determine to what extent the risk of BPD is affected by ventilatory management before the first dose of surfactant. They also found a significant association between low PaCO2 levels prior to the first dose of surfactant and the subsequent development of BPD. In those infants whose lowest PaCO2 level was 29 mm Hg or less, the odds for BPD were 5.6 times those of infants whose lowest PaCO2 level was 40 mm Hg or more. The association of early hypocapnia and BPD was also present in those infants believed to have less severe lung disease. They concluded that the ventilatory strategies used to achieve lower PaCO2 levels produced the volutrauma which contributed to the development of BPD.
The association between hypocapnia in preterm infants and neurodevelopmental status has also been examined. Graziani et al (Graziani 1992) used logistic regression analysis to examine the relationships of perinatal factors, such as mechanical ventilation and hypocapnia, to the subsequent occurrence of neurodevelopmental and neurosonographic abnormalities. This analysis revealed that in the mechanically ventilated infants, extremely low PaCO2 levels (less than 17 mm Hg) during the first three days of life were associated with a significantly increased risk of moderate to severe periventricular echodensity, large periventricular cysts, grade III/IV intracranial haemorrhage, and cerebral palsy. A lowest PaCO2 of greater than 20 mm Hg was not associated with an increased risk of adverse outcomes. All blood gas measurements beyond the third day of life were unrelated to any neurosonographic or neurodevelopmental abnormalities. Further analysis comparing the infants with and without severe hypocapnia did not reveal any clinical factors related to the mechanical ventilation or severity of the respiratory distress that could further define the relationship between lowest PaCO2 and outcome. A similar relationship was described by Fujimoto et al (Fujimoto 1994) who prospectively studied very low birth weight infants whose gestational age was less than 35 weeks. In this population, perinatal complications such as antepartum haemorrhage or severe perinatal asphyxia accounted for less than half of the infants with cystic periventricular leukomalacia (PVL). In the absence of such perinatal complications, there was a strong correlation between severe hypocapnia (PaCO2 less than 20 mm Hg) and cystic PVL. The duration as well as the severity of hypocapnia has also been associated with adverse neurodevelopmental outcome. Calvert et al (Calvert 1987) retrospectively compared preterm infants with cystic PVL with matched controls. They found that infants with cystic PVL had both lower mean PaCO2 levels and longer periods with PaCO2 levels less than 25 mm Hg.
The relationship between neurodevelopmental risk and hypocapnia resulting from treatment with high frequency jet ventilation (HFJV) has recently been reported by Wiswell et al (Wiswell 1996a). They prospectively evaluated premature infants undergoing HFJV with serial neurosonograms, and assessed the cumulative effects of hypotension, acidosis, hypoxaemia and hypocarbia. Using logistic regression analysis, they found that infants with cystic PVL were more likely to have cumulative hypocapnia below a threshold level of 25 mm Hg during the first day of life. However, in a prospective, randomised, controlled trial performed by the same investigators in which HFJV was compared to conventional ventilation in preterm infants with respiratory distress syndrome (Wiswell 1996b), hypocapnia was not found to independently predict an adverse outcome. In this study, treatment with HFJV significantly increased the risk of developing cystic PVL.
In response to the growing concern regarding the association between hypocapnia and increased pulmonary and neurodevelopmental morbidity, the practice of allowing higher PaCO2 levels has developed. The intentional hypoventilation which allows, or specifically targets, high PaCO2 levels is known as permissive hypercapnia (Dries 1995). In this approach, the sometimes aggressive pursuit of normal arterial blood gases, with the subsequent increase in lung injury and possible other neurological morbidities, is avoided, but balanced against the less well known effects of hypercapnia and acidosis. The use of permissive hypercapnic ventilation is becoming widespread in adult critical care (Dries 1995). The effects of hypercapnia on physiological systems, particularly in animal models, have been well summarised (Feihl 1994), and most undesired physiologic effects of moderate respiratory acidosis (PaCO2 <= 80 mm Hg, arterial pH >= 7.15) are thought to be reversible. Two detailed reviews of permissive hypercapnia in acute respiratory failure in adults concluded that experimental and uncontrolled clinical evidence suggests an overall benefit, but there was disagreement as to whether the evidence is convincing enough to warrant its introduction into clinical practice (Tuxen 1994, Bidani 1994). Both did, however, highlight the requirement for randomised, controlled trials.
In early 1998, two randomised trials in adults were simultaneously reported. Amato et al (Amato 1998) compared a strategy of protective ventilation, which involved limitations on both positive end-expiratory pressure and tidal volumes, permissive hypercapnia and the preferential use of pressure-limited ventilatory modes, to conventional ventilation. In the group randomised to the protective strategy, there was improved survival at 28 days and a lower rate of barotrauma. However, protective ventilation was not associated with a higher rate of survival to hospital discharge. A similar but smaller trial by Stewart et al (Stewart 1998) reported that there was no difference in mortality, incidence of barotrauma or multiple organ dysfunction and organ failure in patients randomised to either pressure- and volume limited ventilation with permissive hypercapnia or conventional ventilation. More recently, a large randomised, multicentre trial comparing ventilation with lower tidal volumes compared to traditional tidal volumes for acute lung injury and acute respiratory distress syndrome (ARDS) in adults was reported (ARDS Network 2000). In this trial which targeted specific tidal volumes rather than PaCO2 or pH ranges, there were significant reductions in mortality and number of days without ventilator use in the group treated with lower tidal volumes. The results of this well-conducted study suggest that a ventilation strategy designed to protect the lungs from excessive stretch improve important clinical outcomes in patients with acute lung injury and ARDS.
In the management of persistent pulmonary hypertension of the newborn (PPHN), a condition complicating sepsis or meconium aspiration syndrome or occurring in an idiopathic form in term and near-term infants, hyperventilation may be employed to decrease pulmonary arterial pressure. To achieve a significant effect with a resultant increase in oxygenation, it may be necessary to lower the PaCO2 to levels below 20 mm Hg and to increase the pH to above 7.60 (Drummond 1981), although the response may be more directly related to the increase in pH rather than the decrease in PaCO2. However, hypocapnia resulting from the hyperventilation in PPHN management has been shown to be associated with an increased risk for sensorineural hearing loss (Hendricks-Munoz 1988) and low psychomotor developmental test scores (Ferrara 1984). In order to avoid hyperventilation, Wung et al (Wung 1985) described a technique of "gentle ventilation" in infants with PPHN, aiming for PaCO2 levels up to 60 mm Hg. Although they claimed success in all 15 patients, this study is uncontrolled and long term neurodevelopmental outcome was not reported.
At present, experimental animal and uncontrolled clinical evidence suggests that hyperventilation and hypocapnia may be associated with increased pulmonary and neurodevelopmental morbidity. However, a causal relationship has not been established. On the other hand, the safety of hypercapnic ventilatory strategies in the newborn has also not been established. It remains unknown whether it is better to avoid hyperventilation and hypocapnia, aim for normal PaCO2 levels, or to specifically target higher PaCO2 levels. The "safe" or ideal range for PaCO2 is yet to be determined. If such a range does exist, it may not be the same for all neonates but may vary according to birthweight and/or gestational age. Furthermore, there may be variation in effect dependent upon the ventilatory strategy, eg HFOV versus CMV. This can only be established by large randomised controlled clinical trials which examine clinically important short term benefits as well as long term neurodevelopmental outcomes.
A secondary objective is to assess whether a strategy of permissive hypercapnia is associated with significant side effects.
Sub-group analyses are planned to determine whether the results differ by:
Population:
i. gestational age
ii. birthweight
Intervention:
i. type of ventilation - intermittent positive pressure ventilation
(IPPV) volume or pressure controlled or high frequency ventilation (HFV)
ii. target PaCO2 or pH for permissive hypercapnia - eg. 45-55 mmHg,
55-65 mmHg, >65 mmHg
A mechanical ventilation strategy which allows for a target arterial pH less than the normal range (pH 7.35 to 7.45) or less than a specified higher range of pH.
Immediate adverse effects such as altered haemodynamics or cerebral blood flow, and oxygenation.
Weighted mean differences (WMD) are reported for continuous variables such as duration of oxygen therapy. For categorical outcomes such as mortality, the relative risks (RR) are reported. For significant findings, the risk difference (RD) and number need to treat (NNT) are also reported. The fixed effects model has been used for this review.
Ventilatory Strategy
The definition of permissive hypercapnia, and therefore the ventilatory
strategy employed to achieve the targets, differed slightly between the
two studies. One study (Mariani 1999) aimed
to maintain PaCO2 between 45 and 55 mmHg, and pH > 7.20 in the intervention
(permissive hypercapnia) group, whilst the goal in the control (normocapnia)
group was to maintain PaCO2 between 35 and 45 mmHg and PaCO2 > 7.25. These
goals were used for the first 96 hours following randomisation, after which
time higher levels of PaCO2 were also allowed in the control group with
the ventilatory changes directed at the pH criteria. Changes in the ventilator
settings were made according to a modified clinical algorithm. The data
from 49 randomised infants are available for analysis.
The other included study (Carlo 1999) aimed for a target PaCO2 > 52 mmHg in the intervention group (minimal ventilation) and PaCO2 < 48 mmHg in the control group (routine ventilation). These ventilatory strategies were maintained for 10 days, using a mechanical ventilation protocol. This study was a component of a 2 X 2 factorial design, multicentre, randomised controlled trial of the two ventilatory strategies and early postnatal steroids versus placebo in the first 10 days of life. This trial was terminated early due to unanticipated side effects of steroids. The data from 220 randomised infants are available for analysis.
Outcomes
Both studies reported respiratory and non-respiratory outcomes.
One study (Carlo 1999) reported chronic lung
disease (CLD) using the definition of O2 dependency at 36 weeks postmenstrual
age, and reported pulmonary interstitial emphysema (PIE) and pneumothorax
separately. The other study (Mariani 1999)
reported bronchopulmonary dysplasia (BPD) using the definition of oxygen
requirement and abnormal chest radiograph on day 28, and pneumothorax and/or
PIE were combined and reported as air leak. Both studies reported intraventricular
haemorrhage (IVH) grade 3-4, and periventricular leukomalacia (PVL). Neither
study reported neurodevelopmental outcomes.
Additional data
Contact has been made with authors of both included studies and additional
data has been requested but not yet received. Thus, this review includes
published data only.
Mortality
One trial (Mariani 1999) reported this
outcome. There was no evidence of effect on mortality. (RR 1.04, 95% CI
0.23, 4.66)
Bronchopulmonary Dysplasia and Chronic Lung Disease (in survivors)
One trial (Mariani 1999) reported these
outcomes. There was no evidence of effect on either bronchopulmonary dysplasia
defined as oxygen requirement and abnormal chest radiograph at 28 days
of postnatal age (RR 0.64, 95% CI 0.36, 1.15) or chronic lung disease defined
as oxygen requirement at 36 weeks (RR 1.05, 95% CI 0.16, 6.77).
Death or Chronic Lung Disease at 36 weeks
Both trials (Carlo 1999; Mariani
1999) reported the combined outcome of death or CLD. There was no evidence
of effect in either of the individual trials or in the meta-analysis (RR
0.94, 95% CI 0.78, 1.15).
Intraventricular Haemorrhage and Periventricular Leukomalacia
Intraventricular haemorrhage of all grades was reported in one trial
(Mariani 1999), while both trials reported
IVH grades 3 and 4 (Carlo 1999; Mariani
1999). There was no evidence of effect on IVH all grades (RR 0.82,
95% CI 0.47, 1.43) or IVH grades 3 and 4 (RR 0.84, 95% CI 0.54, 1.31).
Both trials reported PVL (Carlo 1999; Mariani
1999). Neither found evidence of effect, and the meta-analysis did
not support an effect on PVL (RR 1.02, 95% CI 0.49, 2.12).
Air Leak
In the one trial (Mariani 1999) which reported
airleak as a single outcome, there was no evidence of effect (RR 0.52,
95% CI 0.10, 2.59). One trial (Carlo 1999) reported
pulmonary interstitial emphysema and pneumothorax as separate outcomes.
This trial found no evidence of effect on either PIE (RR 1.14, 95% CI 0.63,
2.07) or pneumothorax (RR 2.29, 95% CI 0.73, 7.22).
Retinopathy of Prematurity
One trial (Mariani 1999) reported ROP stage
2 or above. There was no evidence of effect on ROP (RR 1.04, 95% CI 0.07,
15.73).
Duration of Hospital Stay
The duration of hospital stay (days) was reported in one trial (Mariani
1999). There was no evidence of effect on duration of hospital stay
(WMD -2.00, 95% CI -14.87, 10.87).
Subgroup analysis
Due to the limited availability of data, subgroup analyses according
to either population or type of intervention were not possible. However,
one study (Carlo 1999) reported a reduction in
CLD or death in the subgroup of 501 to 750 gram infants (minimal ventilation
strategy 68% vs routine ventilation 86%, p < 0.05) due to a reduction
in CLD in this group. This outcome does not appear in the table of comparisons
as the number of infants in each subgroup was not reported.
The ventilatory targets used in each trial were similar, although the study periods were different. It should be noted that the trial which reported the reduction in CLD in the 501 to 750 gram subgroup (Carlo 1999) maintained the minimal ventilation strategy for a period of 10 days. During this time, the mean PCO2 was 4 mmHg higher than in the intervention group. It is not clear, however, how quickly these targets were achieved. In the trial which maintained the target PCO2 for 96 hours (Mariani 1999), the difference in PCO2 levels did not become significant until 12 hours after randomisation. If permissive hypercapnia/minimal ventilation techniques were to reduce pulmonary morbidity and be neuroprotective, the benefits of these strategies may only be realised if they are instituted immediately following birth, and maintained throughout the duration of the ventilatory support.
Experimental animal data and observational data from human infants (Garland 1995, Graziani 1992, Calvert 1987) suggest that the levels of PCO2 which may be either harmful or protective to the lungs and immature brain are more extreme than those which were targeted in the two included studies. A higher PCO2 and lower pH in the intervention group may be required in order to demonstrate the beneficial effects suggested by the experimental and observational data.
There were no adverse effects due to the permissive hypercapnia/minimal ventilation strategies reported in the included studies. It can be concluded that hypercapnia at least in the range targeted in these two trials is not harmful in the short term. However, long term neurodevelopmental evaluation has not been reported. The safety of prolonged exposure to higher levels of CO2 remains to be shown in controlled studies.
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Carlo 1999 | RCT
2x2 factorial design (with early postnatal steroids) Blinding of randomisation: Can't tell from abstract Blinding of intervention: No Complete followup: Yes Blinding of outcome: Can't tell from abstract |
220 newborn infants.
501-1000 grams birthweight. Mechanically ventilated at <12 hours old. If 751-1000 grams birthweight then FiO2 >= 0.3 and received surfactant. |
Intervention group -
Minimal ventilation with PaCO2 goal of >52 mmHg for 10 days. (n = 109) Control group -
|
CLD or death.
PIE. Pneumothorax. IVH grade 3 or 4. PVL. Reintubation. |
The other arms of the factorial design compared early steroids with
placebo.
The entire study was terminated early because of "unanticipated adverse effects of steroids". |
B |
| Mariani 1999 | RCT
Blinding of randomisation: Yes Blinding of intervention: No Complete followup: Yes Blinding of outcome: Can't tell |
49 newborn infants.
601-1250 grams birthweight. Had RDS and ventilated <24 hours old. Received surfactant. |
Intervention group -
Permissive hypercapnia - PaCO2 45-55 mmHg and pH >= 7.2. (n = 24) Control group - PaCO2 35-45 mmHg and pH >= 7.25. (n = 25) These goals maintained for the first 96 hours following randomisation. After 96 hours the changes in the ventilator settings were directed at the pH criteria allowing high levels of PaCO2 also in the control group. |
Primary -
Duration assisted ventilation. Secondary -
|
A |
Carlo WA, Stark AR, Bauer C, Donovan E, Oh W, Papile L-A, Shankaran S, Tyson JE, Wright LL, Temprosa, Poole K. Effects of minimal ventilation in a multicenter randomized controlled trial of ventilator support and early corticosteroid therapy in extremely low birthweight infants. Pediatrics 1999;104(3, Suppl):738-739.
Mariani 1999 {published data only}
Mariani G, Cifuentes J, Carlo WA. Randomized trial of permissive hypercapnia in preterm infants. Pediatrics 1999;104:1082-1088.
Full report of Carlo 1999 expected in 2001.
* indicates the primary reference for the study
Amato MB, Barbas CS, Medeiros DM, et al. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. New Eng J Med 1998;338:347-354.
The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301-1308.
Avery ME, Tooley WH, Keller JB, et al. Is chronic lung disease in low birth weight infants preventable?: A survey of eight centers. Pediatrics 1987;79:26-30.
Bidani A, Tzouanakis AE, Cardenas VJ, et al. Permissive hypercapnia in acute respiratory failure. JAMA 1994;272:957-962.
Calvert SA, Hoskins EM, Fong KW, et al. Etiologic factors associated with the development of periventricular leukomalacia. Acta Paediatr Scand 1987;76:254-259.
Dreyfuss D, Sauman G. Role of tidal volume, FRC, and end-expiratory volume in the development of pulmonary oedema following mechanical ventilation. Am Rev Respir Dis 1993;148:1194-203.
Dries DJ. Permissive hypercapnia. J Trauma 1995;39:984-989.
Drummond WH, Gregory GA, Heyman MA, et al. The independent effects of hyperventilation, tolazoline, and dopamine on infants with persistent pulmonary hypertension. J Pediatr 1981;98:603-11.
Feihl F, Perret C. Permissive hypercapnia. How permissive should we be? Am J Respir Crit Care Med 1994;150:1722-1737.
Ferrara B, Johnson DE, Change P-N, et al. Efficacy and neurologic outcome of profound hypocapneic alkalosis for the treatment of persistent pulmonary hypertension in infancy. J Pediatr 1984;105:457-61.
Fujimoto S, Togari H, Yamagichi N, et al. Hypocarbia and cystic periventricular leukomalacia in premature infants. Arch Dis Child 1994;71:F107-F110.
Garland JS, Buck RK, Allred EN, et al. Hypocarbia before surfactant therapy appears to increase bronchopulmonary dysplasia risk in infants with respiratory distress syndrome. Arch Pediatr Adolesc Med 1995;149:617-622.
Graziani LJ, Spitzer AR, Mitchell DG, et al. Mechanical ventilation in preterm infants: Neurosonographic and developmental studies. Pediatrics 1992;90:515-522.
Hendricks-Munoz KD, Walton JP. Hearing loss in infants with persistent fetal circulation. Pediatrics 1988;81:650-56.
Kraybill EN, Runyan DK, Bose CL, et al. Risk factors for chronic lung disease in infants with birth weights of 751 to 1000 grams. J Pediatr 1989;115:115-120.
Moylan FM, Walker AM, Kramer SS, et al. Relationship of bronchopulmonary dysplasia to the occurrence of alveolar rupture during positive pressure ventilation. Crit Care Med 1978;6:140-42.
Northway WH Jr, Rosan RC, Porter DY. Pulmonary disease following respiratory therapy of hyaline membrane disease: BPD. N Engl J Med 1967;276:357-368.
Stewart TE, Meade MO, Cook DJ, et al. Evaluation of a ventilation strategy to prevent barotrauma in patients at high risk for acute respiratory distress syndrome. N Engl J Med 1998;338:355-361.
Tuxen DV. Permissive hypercapnic ventilation. Am J Respir Crit Care Med 1994;150:870-874.
Wiswell TE, Graziani LJ, Kornhauser MS, et al. Effects of hypocarbia on the development of cystic periventricular leukomalacia in premature infants treated with high-frequency jet ventilation. Pediatrics 1996;98:918-924.
Wiswell TE, Graziani LJ, Kornhauser MS, et al. High frequency jet ventilation in the early management of respiratory distress syndrome is associated with a greater risk for adverse outcomes. Pediatrics 1996;98:1035-43.
Wung J-T, James LS, Kilchevsky E, et al. Management of infants with severe respiratory failure and persistence of the fetal circulation without hyperventilation. Pediatrics 1985;76:488-494.
Yu VY, Orgill AA, Lim SB, et al. Bronchopulmonary dysplasia in very low birthweight infants. Aust Paed J 1983;19:233-236.