Duration of Mechanical Ventilation in Life-Threatening Pediatric Asthma: Description of an Acute Asphyxial Subgroup
Objective. Acute asphyxial asthma (AAA) is well described in adult patients and is characterized by a sudden onset that may rapidly progress to a near-arrest state. Despite the initial severity of AAA, mechanical ventilation often restores gas exchange promptly, resulting in shorter durations of ventilation. We believe that AAA can occur in children and can lead to respiratory failure that requires mechanical ventilation. Furthermore, children with rapid-onset respiratory failure that requires intubation in the emergency department (ED) are more likely to have AAA and a shorter duration of mechanical ventilation than those intubated in the pediatric intensive care unit (PICU).
Methods. An 11-year retrospective chart review (1991-2002) was conducted of all children who were aged 2 through 18 years and had the primary diagnosis of status asthmaticus and required mechanical ventilation.
Results. During the study period, 33 (11.4%) of 290 PICU admissions for status asthmaticus required mechanical ventilation. Thirteen children presented with rapid respiratory failure en route, on arrival, or within 30 minutes of arrival to the ED versus 20 children who progressed to respiratory failure later in their ED course or in the PICU. Mean duration of mechanical ventilation was significantly shorter in the children who presented with rapid respiratory failure versus those with progressive respiratory failure (29 ± 43 hours vs 88 ± 72 hours). Children with rapid respiratory failure had greater improvements in ventilation and oxygenation than those with progressive respiratory failure as measured by pre- and postintubation changes in arterial carbon dioxide pressure, arterial oxygen pressure/fraction of inspired oxygen ratio, and alveolar-arterial gradient. According to site of intubation, 23 children required intubation in the ED, whereas 10 were intubated later in the PICU. Mean duration of mechanical ventilation was significantly shorter in the ED group versus the PICU group (42 ± 63 hours vs 118 ± 46 hours). There were significantly greater improvements in ventilation and oxygenation in the ED group versus the PICU group as measured by pre- and postintubation changes in arterial carbon dioxide pressure and arterial oxygen pressure/fraction of inspired oxygen ratio.
Conclusions. AAA occurs in children and shares characteristics seen in adult counterparts. Need for early intubation is a marker for AAA and may not represent a failure to maximize preintubation therapies. AAA represents a distinct form of life-threatening asthma and requires additional study in children.
The prevalence of asthma continues to rise, with 5% to 10% of all children estimated to be affected.1 Five percent of these children will be hospitalized during childhood2; of these, an even smaller number require intensive care. Respiratory failure is uncommon, occurring in only 8% to 24% of children who have asthma and are admitted to intensive care units.3–9 Although few children have life-threatening asthma episodes, these episodes are associated with potential mortality, a high morbidity, and a high cost of treatment. These life-threatening episodes can either present as a progressive exacerbation refractory to escalating therapies or have an abrupt onset with rapid deterioration to respiratory failure.
Acute asphyxial asthma (AAA), or rapid-onset near-fatal asthma, is well described in adults. AAA has a predilection for young male adults and is characterized by a brief duration of symptoms (usually <6 hours), few identifiable triggers, and a rapid progression to respiratory failure.10,11 Often, the patient will present in extremis, cyanotic, with little to no air movement, and obtundation.12 Despite the severity of presentation, response to therapy is prompt. When mechanical ventilation is warranted, its duration is usually short, as a result of rapid improvements in gas exchange.12,13
Near-fatal asthma episodes are known to occur in children.14–17 There seems to be a subset of children with near-fatal episodes whose presentation and clinical course may be comparable to those seen in AAA in adults.18
We hypothesized that children who have rapidly progressive asthma and require early mechanical ventilation will have more rapid improvements in gas exchange and require shorter durations of mechanical ventilation than children who have asthma and slowly progress to respiratory failure. The need for early intubation and a shorter duration of mechanical ventilation may serve as markers of AAA in these children.
We identified and reviewed all charts of patients who were between 2 through 18 years of age and had a primary discharge diagnosis of asthma and required mechanical ventilation at Golisano Children's Hospital at Strong (GCHaS) during an 11-year period (1991-2002). Children who were younger than 2 years were excluded because of the significant overlap in the diagnoses of asthma and bronchiolitis. The pediatric intensive care unit (PICU) at GCHaS is a 12-bed medical-surgical unit with ∼650 annual admissions. It serves as the regional referral center in central western New York State.
Before chart review, approval was obtained from the GCHaS institutional review board. Two authors (F.M., S.S.) independently reviewed all charts using a standardized approach. Children with other preexisting respiratory diseases (eg, bronchopulmonary dysplasia, interstitial lung disease) or significant comorbidities (eg, congenital heart disease, malignancy, sickle cell disease) were excluded from the review.
The following data were collected (when available): patient demographics, asthma history, preexacerbation therapies, duration of symptoms before exacerbation, exacerbation therapies (including mechanical ventilation), duration of mechanical ventilation, and results from arterial blood gases drawn immediately before and after intubation.
Patients were grouped and analyzed in 2 ways: 1) time of onset of respiratory failure and 2) site of intubation. By time of onset of respiratory failure, we compared children with rapid respiratory failure (RRF) with those with progressive respiratory failure (PRF). RRF was defined as respiratory failure diagnosed en route, on arrival to the emergency department (ED), or within 30 minutes of arrival to the ED. We defined respiratory failure as the need for assisted ventilation as a result of cardiopulmonary arrest, apnea, central cyanosis despite high flow oxygen, or progressive obtundation.
When grouping children by site of endotracheal intubation and initiation of mechanical ventilation, we compared clinical characteristics of children who were intubated in the ED with those who were intubated in the PICU. We chose this secondary comparison to attempt to detect a difference in the threshold for the initiation of mechanical ventilation on the basis of alternative practice approaches in the ED versus the PICU.
For assessing differences between groups, [chi]2 analyses were applied to categorical data and unpaired t tests were applied to interval data. P < .05 was used as the criterion for rejection of the null hypothesis. A commercially available statistical program (Microsoft Excel 2000, Redmond, WA) was used for all calculations.
During the 11-year study period, there were 290 admissions to the PICU for severe asthma. Of these admissions, 33 (11.4%) required mechanical ventilation. The 33 episodes of mechanical ventilation occurred among 32 children (12 girls, 20 boys); 1 girl required mechanical ventilation on 2 occasions. The mean age of the study group was 11.8 years with a range of 2 to 18 years. There were no deaths among the study group. No patient had a pneumothorax before intubation or on the chest radiograph obtained immediately after intubation.
Comparison by Time of Onset of Respiratory Failure
Thirteen children had RRF, and 20 children had PRF. Of the 13 children who presented with RRF, 4 required bag mask ventilation or intubation en route by emergency medical services as a result of profound cyanosis, apnea, or cardiopulmonary arrest. Two additional children were intubated immediately on arrival for obvious respiratory failure, and 7 children were intubated within 30 minutes of arrival (mean: 8.5 minutes) as a result of respiratory failure during initial treatment. All 7 of these children were receiving a high-dose continuous inhalation of a β-agonist at the time of respiratory failure.
Of the 20 children with PRF, 10 developed respiratory failure that required intubation late in their ED course (34-120 minutes after arrival; mean: 72 minutes; median: 68 minutes) and 10 developed respiratory failure that required intubation after transfer to the PICU (2-24 hours after hospital arrival; mean: 9 hours 6 minutes; median: 5 hours 35 minutes).
Children with RRF were older than those with PRF (13.3 ± 2.3 years vs 10.8 ± 3.9 years; P = .03). There were no significant differences in gender, race, previous PICU admissions or intubations, or maintenance asthma medications in the RRF group versus the PRF group. No trigger for the asthma attack was identified in 8 (62%) of 13 children with RRF versus 9 (45%) of 20 children with PRF. This difference was not statistically significant (P = .35)
Children with RRF had a significantly shorter mean duration of mechanical ventilation than those with PRF (29.2 ± 43 hours vs 88.2 ± 72 hours; P = .006; Fig 1). We also analyzed the data post hoc using cutoffs of 5 (intubated on arrival or en route), 90, and 180 minutes; in all cases, the difference between groups in duration of ventilation remained significant, but the level of significance decreased (P = .004, .01, and .02, respectively).
Paired preintubation (<30 minutes before intubation) and postintubation (<30 minutes after intubation) arterial blood gas data were available in 6 of 13 children with RRF and in 14 of 20 children with PRF. Although there were no differences between groups in immediate preintubation oxygen saturation (RRF 71% vs PRF 88%; P = .12), preintubation arterial oxygen pressure (Pao2)/fraction of inspired oxygen (Fio2) ratio, preintubation alveolar-arterial (A-a) gradient, or preintubation partial pressure of carbon dioxide (Paco2), there were significant differences between groups in improvements of gas exchange as measured by changes in Paco2, Pao2/Fio2 ratio, and A-a gradients (Tables 1 and 2). The RRF group had a mean reduction in Paco2 of 46 mm Hg versus a 1-mm Hg increase in the PRF group (P = .01). The RRF group had a mean improvement in Pao2/Fio2 ratio of 260 versus 88 in the PRF group (P = .04). Similarly, the RRF group had a mean decline of 327 in the A-a gradient versus a decline of 100 in the PRF group (P = .006).
Comparison by Site of Intubation
Twenty-three children were intubated the ED versus 10 children in the PICU. There were no significant differences between groups in age (mean: 12.1 years vs 11.0 years), gender, race, previous PICU admissions or intubations, or maintenance asthma medications. Children who were intubated in the ED had a shorter mean duration of mechanical ventilation (41.9 ± 63 hours vs 118 ± 46 hours in children who were intubated in the PICU; P = .0008; Fig 2).
Paired pre- and postintubation arterial blood gas data were available in 12 of 23 children who were intubated in the ED and in 8 of 10 children who were intubated in the PICU. Although there were no differences between groups in immediate preintubation oxygen saturations (ED 80.6% vs PICU 84.1%; P = .66), preintubation Pao2/Fio2 ratio, preintubation A-a gradient, or preintubation Paco2, there were significant differences in improvement in ventilation and oxygenation between groups (Tables 3 and 4). The ED group had a mean reduction in Paco2 of 35 mm Hg versus a 21-mm Hg increase in the PICU group (P = .004) and a mean improvement in Pao2/Fio2 ratio of 217 versus 43 in the PICU group (P = .02).
We found no differences in the duration of symptoms before presentation when comparing the RRF and the PRF groups (35 hours vs 30 hours; P = .61). Likewise, there were no differences in the duration of symptoms when comparing the ED group and the PICU group (ED 32.4 hours vs PICU 31.8 hours; P = .95). We did note, however, a trend toward shorter lengths of ventilation among children with duration of symptoms of 6 hours or less. Among children who were intubated in the ED, 9 with short symptom duration (<6 hours) had a mean length of ventilation of 24 hours versus 54 hours in children with longer symptom duration (P = .1).
Using early intubation and a brief duration of mechanical ventilation as clinical proxies, we have identified a population of children who are likely to have a disease process similar to AAA seen in adults. Although several authors have described adolescents with fatal and near-fatal AAA events,14–17 descriptions of AAA events in younger children are lacking. An Australian study identified 30 near-fatal asthma episodes that occurred in children who were younger than 15 years. Most cases were marked by progressive respiratory distress, but 5 (17%) had sudden onset with rapid respiratory collapse.18 Although we found that children with RRF tended to be older than those with PRF, a significant portion of children in the present study who had RRF were preadolescent; 4 (31%) of the 13 children with RRF were 12 years or younger (age range: 10-18 years).
We chose to attempt to identify children with AAA using the objective finding of respiratory failure evident on presentation or shortly thereafter (during the ED course). Although prospective determinants of respiratory failure could not be used, evidence of true respiratory failure was clear in all cases: requirement of assisted ventilation as a result of cardiopulmonary arrest, apnea, central cyanosis despite high flow oxygen, or progressive obtundation. Arterial blood gas data were not obtained in many of these overt cases because of the need to initiate therapy promptly. In these cases, patients necessarily received a diagnosis of respiratory failure on clinical grounds; obtaining blood gases would have been injudicious and posed needless risk to these patients. When blood gas data were available, we used these data to examine improvements in gas exchange quantitatively.
We chose intubation within 30 minutes of ED arrival as a cutoff time in delineating RRF from PRF. Although arbitrary, 30 minutes is a rough estimate of the time for initiation of urgent “rescue” therapy (nebulized, subcutaneous, and/or intravenous β-agonists). Children who remained in respiratory failure (or deteriorated further) during this brief trial period without mechanical support did have clinical findings and blood gas data (when available) supporting a rapidly progressive disease process.
We found that children who had asthma with respiratory failure on presentation or shortly thereafter often had prompt normalization of gas exchange and brief durations of mechanical ventilation. This is consistent with adult descriptions of AAA. Plaza et al13 prospectively studied 220 adults with near-fatal asthma attacks and identified 45 as rapid-onset and rapidly progressive attacks. Patients in the rapid-onset group often presented in arrest and had very brief durations of ventilation when compared with the slow-onset group. Likewise, Wasserfallen et al12 found rapid recovery and short durations of ventilation (mean: 34 hours vs 91 hours) in adults who were characterized as having sudden asphyxial asthma compared with those having a more gradual development of respiratory failure.
An expected shorter duration of symptoms before intubation was not seen in our children with RRF as compared with those who developed respiratory failure later. This may be explained by several limitations to our study. First, the retrospective nature of the study does not allow for standardized documentation of historical data. The recorded onset and duration of symptoms before presentation may have been inexact. Historical documentation of symptom duration was often approximated, varied according to clinician (eg, resident note vs attending note), and, at times, was lacking. Documentation of the onset of symptoms was variable. Some historians documented the duration as beginning with any symptom of an upper respiratory infection (eg, rhinorrhea, cough), whereas others considered wheezing or chest discomfort as the onset. Therefore, grouping children according to duration of symptoms (rapid onset vs slow onset) could lead to inaccurate conclusions about differences between groups. Last, our study may have lacked the number of patients necessary to detect such a difference.
Another limitation of the study was lack of standardized preintubation treatment strategies. The lack of a uniform treatment protocol may have had an impact on the timing of the initiation of mechanical ventilation. This would have a greater importance in children who were intubated later in their hospital course versus those who presented in respiratory failure. Reviewing the treatment of the 20 children who developed respiratory failure later in their hospital course, we found that all children were receiving high-dose continuous inhalation β-agonist (albuterol 15-20 mg/hour) and intravenous steroids (methylprednisolone 2-4 mg/kg/day) before intubation. Additional therapies before intubation were used as follows: intravenous β-agonist (terbutaline 0.5-5.0 μg/kg/min) in 18 (90%) of 20, inhaled ipratropium bromide or atropine in 19 (95%) of 20, intravenous magnesium sulfate in 10 (50%) of 20, and provision of a heliox gas mixture (70% helium:30% oxygen) in 6 (30%) of 20. No child had a trial of noninvasive ventilation attempted. Of note, the rate of mechanical ventilation among children who had asthma and were admitted to our PICU (11.4%) during the study period was similar to or lower than that reported previously.3–6
AAA events may have a unique pathophysiology; specifically, intense bronchospasm with a relative absence of bronchial inflammation. The contribution of mucus plugging to an AAA episode is unclear. There have been clinical reports of a paucity of mucus in adults with AAA during routine suctioning12,16; however, postmortem examination of adults who have died of AAA has revealed an increased mucus gland area.19 Several authors have described a characteristic submucosal cellular profile in adults who have died during an AAA event. Surprising, neutrophils are the predominate cell type in cases of AAA, whereas eosinophils predominant in slow-onset cases.19–21 Several theories have been postulated to explain these cellular differences. Neutrophil predominance in AAA may simply be a function of time. In animal models, the recruitment of neutrophils by a specific stimulus may precede and, in fact, lead to later eosinophilic infiltration.22,23 Second, the nature of the stimulus itself may lead to a specific initial cellular response. Bacterial endotoxin and certain environmental agents such as ozone have been shown to cause neutrophil-mediated bronchial hyperresponsiveness.24,25 Last, neutrophils may not have been a primary component of the AAA event and indeed were “innocent bystander” cells.20 An initial neurogenic event may have mediated intense bronchospasm, independent of the submucosal cellular profile. Alternatively, the predominance of bronchial hyperresponsiveness with a relative lack of bronchial inflammation may be a function of an early therapeutic intervention (mechanical ventilation) interrupting a pathophysiologic sequence of events. This seems unlikely, as mechanical ventilation has been known to induce proinflammatory lung changes.26,27
In conclusion, we believe that AAA can occur in children and can lead to early respiratory failure. Our patients with RRF caused by asthma share many key characteristics seen in adults, namely, severity of presentation, rapid improvements in gas exchange after initiation of mechanical ventilation, and subsequent brief durations of mechanical ventilation. As in adults, AAA in children may represent a pathologically distinct form of asthma. Therefore, the need for early intubation and a short duration of mechanical ventilation may serve as proxies for AAA and may not represent a failure to maximize preintubation therapies. Prospective study will be required to clarify further these hypotheses.
- ↵Mannino DM, Homa DM, Akinbami LJ, et al. Surveillance for asthma—United States, 1980–1999. MMWR CDC Surveill Summ.2002;51 :1– 6
- ↵Kolbe J, Fergusson W, Garrett J. Rapid onset asthma: a severe but uncommon manifestation. Thorax.1998;53 :241– 247
- ↵Plaza V, Serrano J, Picado C, et al. Frequency and clinical characteristics of rapid-onset fatal and near-fatal asthma. Eur Respir J.2002;19 :846– 852
- Saetta M, Thiene G, Crescioli S. Fatal asthma in a young patient with severe bronchial hyperresponsiveness but stable peak flow records. Eur Respir J.1989;2 :1008– 1012
- ↵Schmitz T, von Kries R, Wjst M, et al. A nationwide survey in Germany on fatal and near-fatal asthma in children: different entities? Eur Respir J.2000;16 :845– 849
- ↵Carrol N, Carello S, Cooke C, et al. Airway structure and inflammatory cells in fatal attacks of asthma. Eur Respir J.1996;9 :709– 715
- ↵Hunt LW, Mansfield ES, Sur S, et al. Late neutrophilic response to bronchial allergen challenge: a response to endotoxin? J Allergy Clin Immunol.1992;89 :335
- ↵Dos Santos CC, Slutsky AS. Mechanisms of ventilator induced lung injury: a perspective. J Appl Physiol.2000;89 :1645– 1655
- Copyright © 2004 by the American Academy of Pediatrics