Anesth Analg 2006;103:1219-1223
© 2006 International Anesthesia Research Society
doi: 10.1213/01.ane.0000237433.00877.5a
CRITICAL CARE AND TRAUMA
Initial and Delayed Onset of Acute Respiratory Failure: Factors Associated with Development and Outcome
Suzana M. Lobo, MD,
Francisco R. M. Lobo, MD,
Flavio Lopes-Ferreira, MD,
Daliana Peres Bota, MD,
Christian Melot, MD, PhD, and
Jean-Louis Vincent, MD, PhD
From the Department of Intensive Care, Erasme Hospital, Free University of Brussels, Brussels, Belgium.
Address correspondence and reprint requests to Jean-Louis Vincent, MD, PhD, Department of Intensive Care Medicine, Erasme University Hospital, Route de Lennik 808, 1070 Brussels, Belgium. Address e-mail to jlvincen{at}ulb.ac.be.
 |
Abstract
|
|---|
In a prospective observational study of 1038 adult admissions to a 31-bed medical/surgical intensive care unit (ICU), acute respiratory failure (ARF, defined as a Pao2/Fio2 ratio 
200 mm Hg and the need for respiratory support) occurred in 182 (58%) of the 313 admissions with an ICU stay of more than 48 h. Initial ARF (onset within 48 h of ICU admission) occurred in 133 (42%) patients, and delayed onset ARF (onset >48 h after ICU admission) in 49 (16%). On admission, the cardiovascular sequential organ failure assessment (SOFA) score was higher in initial than in delayed onset ARF (1.1 ± 1.5 vs 0.6 ± 1.2, P < 0.05). High admission serum C-reactive protein concentrations (OR 1.08, 95% CI 1.04–1.12, P = 0.0001) and SOFA scores (OR 1.20, 95% CI 1.08–1.33, P = 0.0007) were the factors independently associated with initial ARF, and a low Glasgow coma scale (GCS) score (OR 1.13, 95% CI 1.04–1.21, P = 0.0018) was associated with delayed onset ARF. In initial ARF, a high SOFA score (OR 1.24, 95% CI 1.12–1.38, P = 0.0001) and a low GCS score (OR 0.89, 95% CI 0.83–0.96, P = 0.0013) on admission, and in delayed onset ARF, a low GCS score at 48 h (OR 0.67, 95% CI 0.54–0.84, P = 0.0011) were independently associated with death. The mortality rate was similar for initial and delayed onset ARF.
|
Introduction
|
|---|
Acute respiratory failure (ARF) is caused by many factors and remains one of the most common reasons for admission to the intensive care unit (ICU). Despite all the advances of modern medicine, ARF is still associated with mortality rates ranging from 30% to 45% (1–5). With few exceptions (6), studies on ARF usually have not considered the time of onset of ARF, which may influence outcome. Croce et al. (6) reported that in trauma patients, early onset of acute respiratory distress syndrome was associated with severe hemorrhagic shock and capillary leak, whereas late onset acute respiratory distress syndrome often followed infection and was associated with multiple organ failure. Other authors reported the severity of fractures and shock as determinant factors of early ARF in trauma patients (7). Rady et al. (8) reported that early pulmonary dysfunction after cardiovascular surgery was associated with a postoperative increase in serum creatinine, neurological complications, nosocomial infection, and death.
Initial and delayed presentations of organ dysfunction in the ICU may have different features, as has been reported for acute renal failure, with a higher mortality in patients with late onset renal failure compared with those with early onset renal failure (9). Importantly, if the time of onset of organ dysfunction is indeed associated with outcome, this factor should be considered in clinical studies evaluating new forms of therapy, as the timing of therapeutic interventions may also influence outcome.
The aim of the present study was to prospectively evaluate the occurrence, clinical outcome, and predictive factors for survival, for initial and delayed onset ARF in critically ill adult patients.
|
METHODS
|
|---|
As this was an observational study, requiring no deviation from standard care, the local ethics committee waived the need for informed consent. All patients staying in the mixed surgical–medical, 31-bed University Hospital Department of Intensive Care for more than 48 h during a 4-mo period were prospectively evaluated. A database was created including demographic, physiological, and inflammatory variables, use of vasoactive drugs, presence of infection on admission, duration of mechanical ventilation, extracorporeal renal support, and organ function evaluation. Some of the patients were also enrolled in other epidemiological studies (10,11). Infection was determined according to usual clinical, laboratory, and microbiological variables (12). Organ function was evaluated on a daily basis, using the most abnormal value each day, for each of six organ systems (respiratory, renal, cardiovascular, hepatic, coagulation, and neurological) according to the sequential organ failure assessment (SOFA) score (13). When patients were sedated, the Glasgow Coma Scale (GCS) score recorded before sedation was used. If sedation was stopped and a new value for GCS was recorded, this new value was then used. The Acute Physiology and Chronic Health Evaluation II (APACHE II) score (14) was calculated on admission. Serum C-reactive protein (CRP) concentrations were measured on admission and on Day 2 using an immunochemistry analyzer (Modular, Hitachi, Tokyo, Japan). For mean arterial blood pressure (MAP), the lowest value observed during a 24-h period was noted from the patient's records. ARF was defined as a Pao2/Fio2 ratio 200 mm Hg and the need for any form of respiratory support, with or without an artificial airway. Conventional modes of mechanical ventilation (either volume-control or pressure-control ventilation) were used. Patients with ARF were classified according to the time of onset of ARF: initial ARF was defined in patients who met the criteria within 48 h of ICU admission, and delayed onset ARF in patients who met the criteria 48 h or more after ICU admission.
The variables on ICU admission that were evaluated as potentially associated with early ARF were age, gender, APACHE II score, use of vasoactive drugs, serum CRP concentrations, serum lactate concentrations, GCS, presence of infection, platelet count, MAP, total SOFA score, and respiratory, renal, cardiovascular, hepatic, coagulation, and neurological SOFA scores. These same variables, together with the Pao2/Fio2 ratio taken on admission and on Day 2, were evaluated as potentially associated with delayed onset ARF and death.
Results are expressed as mean ± sd. Continuous variables were compared using Student's t-test or analysis of variance (ANOVA) for repeated measurements. Proportions were compared using the Z-test. Univariate logistic regression analysis was used to detect significant predictors of ARF and ICU mortality. The statistically significant factors associated with ARF or ICU mortality were entered into a stepwise multiple logistic regression analysis. For these analyses, a P value <0.05 was considered as statistically significant. Odds ratio (OR) and 95% confidence interval (CI) were calculated from the coefficients and standard errors obtained from the logistic regression.
|
RESULTS
|
|---|
During the study period, 1038 patients were admitted to our ICU; of these, 313 patients stayed in the ICU for more than 48 h. Initial ARF was present in 133 of the 313 patients (42%) and delayed onset ARF in 49 (16%). The overall mortality in ARF patients was 33% (32% initial ARF and 36% delayed onset ARF; OR: 0.79, CI 0.40–1.57, NS), and 10% in those without ARF (initial ARF versus no ARF, RR: 3.18, CI 1.7–5.6; delayed onset ARF versus no ARF, RR: 3.7, CI 1.9–6.9, both P < 0.05). When compared with the patients without ARF, patients with initial and delayed onset ARF had higher SOFA and APACHE scores, longer periods of ventilatory and artificial renal support, longer ICU stays, and a higher incidence of infection present on admission. As expected, initial ARF patients had lower Pao2/Fio2 ratios on admission and more days on mechanical ventilation than patients with delayed onset ARF (Table 1).
Patients with initial ARF had a significantly higher respiratory SOFA score than patients with delayed onset ARF. Initial and delayed onset ARF patients had higher organ scores, other than the respiratory score, compared with patients without ARF (Fig. 1). On admission, the cardiovascular SOFA score was higher in initial than in delayed onset ARF (1.09 ± 1.45 vs 0.61 ± 1.17, P < 0.05), and the neurological score was higher in delayed onset than in initial ARF (1.23 ± 1.57 vs 0.86 ± 1.34, P = 0.12).
View larger version (15K):
[in this window]
[in a new window]
|
Figure 1. Pattern of organ dysfunction on admission, in initial and delayed onset acute respiratory failure (ARF) (organ scores as assessed by the sequential organ failure assessment [SOFA] score). *P < 0.05 vs no ARF; #P < 0.05 vs initial ARF.
|
|
The variables most significantly associated with development of ARF by univariate analysis were the APACHE II score, the SOFA score, the cardiovascular SOFA score, CRP concentrations, low MAP, presence of infection on admission, and use of vasoactive drugs. A multiple logistic regression analysis (Table 2) retained CRP concentrations and the SOFA score as major variables independently and significantly associated with initial ARF.
View this table:
[in this window]
[in a new window]
|
Table 2. Variables Independently Associated with Initial and Delayed Onset Acute Respiratory Failure (ARF) (Multiple Logistic Regression Analysis)
|
|
The variables on admission most significantly associated with development of late ARF were the Pao2/Fio2 ratio, GCS, and the neurological SOFA score. Only the GCS score on admission was retained as independently associated with delayed onset ARF (Table 2).
The variables associated with death were the APACHE II score, GCS, CRP, total SOFA score, neurological SOFA score, cardiovascular SOFA score, renal SOFA score, respiratory SOFA score, coagulation SOFA score, and the Pao2/Fio2 ratio. Independent factors associated with mortality for initial and delayed onset ARF are shown in Table 3.
View this table:
[in this window]
[in a new window]
|
Table 3. Variables Independently Associated with Death in Initial and Delayed Onset Acute Respiratory Failure (ARF) (Multiple Logistic Regression Analysis)
|
|
|
DISCUSSION
|
|---|
In this prospective study, we sought to identify different patterns of organ dysfunction and factors associated with initial and delayed onset ARF. We also aimed to identify the factors associated with death in initial and delayed onset ARF. We found that cardiovascular dysfunction on admission was more frequently associated with initial ARF, and neurological dysfunction on admission was more frequently associated with delayed onset ARF. Importantly, we found that abnormalities in pulmonary physiology early in the course of ARF were of minimal or no prognostic value, but the extent of associated nonpulmonary organ dysfunction as assessed by the SOFA score, and the presence of an altered mental status, were associated with mortality.
ARF is common in the ICU, and occurred in 58% of all patients in our study, with the majority (42%) presenting early. Patients with ARF had high infection and mortality rates, required mechanical ventilation and renal extracorporeal support for prolonged periods, and had higher severity-of-illness scores than non-ARF ICU patients. ARF is a life-threatening condition that often requires ventilatory support. More prolonged periods on ventilatory support were observed for early ARF. It is possible that in initial ARF, the respiratory system had a higher degree of injury, thus requiring more prolonged periods of ventilatory support and recovery than in delayed onset ARF. The higher frequency of infection and cardiovascular system dysfunction observed in patients with initial ARF is suggestive of major injury. Infection was present in 61% and 47% of the patients with initial and delayed onset ARF, respectively.
High SOFA scores on ICU admission were independently associated with development of initial ARF and outcome in our study, showing that the number of extrapulmonary factors is very important. Also important, high levels of CRP were associated with the development of initial ARF. Several authors have reported that CRP, an acute-phase protein, is a marker of inflammation and infection (15–17), and may be associated with an increased risk of organ failure and death (11). This may reflect the frequent prevalence of systemic, as well as pulmonary, infection in patients with ARF. However, it is possible that CRP may be more than an indicator of inflammation in ARF, because experimental studies suggest that CRP may have protective effects during ARF (18–20).
Procalcitonin (PCT) is another commonly used marker of infection and outcome that we could have used as a marker (21,22). However, as for CRP, PCT elevations occur in several noninfectious inflammatory conditions, such as trauma and surgery. We chose to measure CRP rather than PCT in this heterogeneous population of critically ill patients, mainly because CRP, but not PCT, levels were measured routinely in our patients at the time of the study.
A low GCS score on admission was independently associated with delayed onset ARF and with death in initial ARF. A persistently low GCS score after 48 h was associated with death in delayed onset ARF patients. In the ICU, a myriad of critical care conditions can lead to an alteration in mental status, including central nervous system (CNS) lesions, hepatic or renal dysfunction, endocrine diseases, meningitis, osmolar and metabolic disturbances, nutritional deficiencies, and medication, particularly sedation. Sedation is a very common reason for low GCS scores in the ICU. Patients' requirements for sedation vary substantially during supportive interventions and may contribute to prolonged mechanical ventilation and complications (23). A number of factors associated with neurological impairment can increase the risk of ARF, including the presence of tracheal and esophageal tubes, which increase gastro-esophageal reflux and the risk of aspiration of gastric contents (24,25). A so-called CNS-lung axis involving the release of mediators after CNS injuries has been suggested as well (26). Gastroparesis is common in patients with severe traumatic brain injuries and in postoperative patients, and may play a role in increasing the risk of bronchoaspiration (27). The possible role of aspiration of gastric contents in the etiology of late ARF should be considered in patients with neurological compromise. Simple interventions such as semirecumbent body position (25), rigorous control of residual gastric volume, and use of prokinetic drugs (27) may help reduce aspiration.
The oxygenation index in this study was a weak prognostic indicator in patients with ARF. Similar findings have been reported in several other studies (3,28–31), suggesting that patients who die do not die from ARF, but with ARF. Multiple organ dysfunction as assessed by the SOFA score was the best predictor of the development of early ARF and subsequent mortality, regardless of the causative factor. In univariate analysis, the prognostic indices APACHE II and SOFA score (as well as the individual organ scores used in SOFA), low GCS scores, high CRP level, and low Pao2/Fio2 ratio were associated with death. By multivariate analysis, only the SOFA score and the GCS were independently associated with mortality. Vincent et al. (4) also found that the presence of multiple organ failure was an important negative prognostic indicator, and Lopes Ferreira et al. (10) reported that, in a heterogeneous population of critically ill patients, the sequential assessment of organ dysfunction during the first few days of ICU admission provides a good indicator of outcome. In 529 patients with acute respiratory failure, Flaatten et al. (32) reported overall hospital mortality rates of 32.9%; in patients with ARF and no other organ failure, mortality rates were just 14.7%, stressing the importance of concomitant multiple organ failure for prognosis in patients with ARF. A daily follow-up of the nonpulmonary organ dysfunction using the SOFA score, and particularly the persistence of neurological dysfunction at 48 h, may be helpful to evaluate the risk of death in patients with ARF in the ICU.
Despite there being no differences in mortality rates, initial and delayed onset ARF have different features that should be considered in future studies. For instance, antiinflammatory drugs may be more appropriately tested in initial onset ARF due to the frequent prevalence of infection and cardiovascular dysfunction and higher CRP levels, suggesting major inflammation in this phase in comparison with patients with delayed onset ARF. The presence of nonpulmonary organ dysfunction as a factor increasing the risk of death is another important issue in clinical trials in ARF.
We acknowledge that the small number of patients from a single center, and the short study period of only 4 mo with potential seasonal effects, place some limitations on this analysis. In addition, the use of the Pao2/Fio2 ratio as a predictor of mortality without detailed information on ventilatory support and respiratory mechanics is prone to error, as the application of even limited levels of positive pressure can markedly affect oxygenation. However, our findings on frequency as well as the importance of associated multiple organ failure and mental status are in accordance with a large multinational and multicenter study that included 1449 patients (4). Etiology and classification of the cause of ARF can be unclear at this time point, especially in the setting of a general ICU and, if included in the multivariable analysis, may have outweighed the more simple variables used. A systemic inflammatory response is, however, the common pathway for all causes of ARF, so we chose to consider an inflammatory marker as a valuable variable for evaluation soon after ICU admission.
|
CONCLUSIONS
|
|---|
In the presence of initial and delayed onset ARF, mortality is related more to the presence of multiple organ failure than to initial lung function. The extent of associated nonpulmonary organ dysfunction, as assessed by the SOFA score and the presence of an altered mental status, were the best predictors of death in patients with initial ARF. Neurological dysfunction at ICU admission is associated with delayed onset ARF, and persistent neurological dysfunction at 48 h is associated with a poor outcome in these patients. We conclude that monitoring of nonpulmonary organ function in patients with ARF may thus be useful in assessing outcome.
|
Footnotes
|
|---|
Accepted for publication June 29, 2006.
|
REFERENCES
|
|---|
- Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994;149:818–24.[Abstract]
- Vasilyev S, Schaap RN, Mortensen JD. Hospital survival rates of patients with acute respiratory failure in modern respiratory intensive care units. An international, multicenter, prospective survey. Chest 1995;107:1083–8.[Abstract/Free Full Text]
- Luhr OR, Antonsen K, Karlsson M, et al; The ARF Study Group. Incidence and mortality after acute respiratory failure and acute respiratory distress syndrome in Sweden, Denmark, and Iceland. Am J Respir Crit Care Med 1999;159:1849–61.[Abstract/Free Full Text]
- Vincent JL, Akca S, de Mendonca A, et al. The epidemiology of acute respiratory failure in critically ill patients. Chest 2002;121:1602–9.[Abstract/Free Full Text]
- Guerin C, Gaillard S, Lemasson S, et al. Effects of systematic prone positioning in hypoxemic acute respiratory failure: a randomized controlled trial. JAMA 2004;292:2379–87.[Abstract/Free Full Text]
- Croce MA, Fabian TC, Davis KA, Gavin TJ. Early and late acute respiratory distress syndrome: two distinct clinical entities. J Trauma 1999;46:361–6.[ISI][Medline]
- Modig J, Hedstrand U, Wegenius G. Determinants of early adult respiratory distress syndrome. A retrospective study of 220 patients with major fractures. Acta Chir Scand 1985;151:413–8.[ISI][Medline]
- Rady MY, Ryan T, Starr NJ. Early onset of acute pulmonary dysfunction after cardiovascular surgery: risk factors and clinical outcome. Crit Care Med 1997;25:1831–9.[ISI][Medline]
- Guerin C, Girard R, Selli JM, et al.; Rhone-Alpes Area Study Group on acute renal failure. Initial versus delayed acute renal failure in the intensive care unit. A multicenter prospective epidemiological study. Am J Respir Crit Care Med 2000;161:872–9.[Abstract/Free Full Text]
- Lopes Ferreira F, Peres Bota D, et al. Serial evaluation of the SOFA score to predict outcome. JAMA 2001;286:1754–8.[Abstract/Free Full Text]
- Lobo SM, Lobo FR, Bota DP, et al. C-reactive protein levels correlate with mortality and organ failure in critically ill patients. Chest 2003;123:2043–9.[Abstract/Free Full Text]
- Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128–40.[ISI][Medline]
- Vincent JL, de Mendonça A, Cantraine F, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicentric, prospective study. Crit Care Med 1998;26:1793–800.[ISI][Medline]
- Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818–29.[ISI][Medline]
- Póvoa P, Almeida E, Moreira P, et al. C-reactive protein as an indicator of sepsis. Intensive Care Med 1998;24:1052–6.[ISI][Medline]
- Yentis SM, Soni N, Sheldon J. C-reactive protein as an indicator of resolution of sepsis in the intensive care unit. Intensive Care Med 1995;21:602–5.[ISI][Medline]
- Smith RP, Lipworth BJ, Cree IA, et al. C-reactive protein. A clinical marker in community-acquired pneumonia. Chest 1995;108:1288–91.[Abstract/Free Full Text]
- Abernathy VJ, Webster RO, Dahms TE. C-reactive protein inhibits increased pulmonary vascular permeability induced by fMLP in isolated rabbit lungs. Am J Physiol 1996;271:H507–13.
- Heuertz RM, Webster RO. Role of C-reactive protein in acute lung injury. Mol Med Today 1997;3:539–45.[ISI][Medline]
- Casals C, Arias-Diaz J, Valino F, et al. Surfactant strengthens the inhibitory effect of C-reactive protein on human lung macrophage cytokine release. Am J Physiol Lung Cell Mol Physiol 2003;284:L466–72.[Abstract/Free Full Text]
- Ugarte H, Silva E, Mercan D, et al. Procalcitonin as a marker of infection in the intensive care unit. Crit Care Med 1999;27:498–504.[ISI][Medline]
- Assicot M, Gendrel D, Carsin H, et al. High serum procalcitonin concentration in patients with sepsis and infection. Lancet 1993;341:515–8.[ISI][Medline]
- Schweickert WD, Gehlbach BK, Pohlman AS, et al. Daily interruption of sedative infusions and complications of critical illness in mechanically ventilated patients. Crit Care Med 2004;32:1272–6.[ISI][Medline]
- Ibanez J, Penafiel A, Marse P, et al. Incidence of gastroesophageal reflux and aspiration in mechanically ventilated patients using small-bore nasogastric tubes. JPEN J Parenter Enteral Nutr 2000;24:103–6.[Abstract]
- Drakulovic MB, Torres A, Bauer TT, et al. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999;354:1851–8.[ISI][Medline]
- de Oliveira GG, Antonio MP. Role of the central nervous system in the adult respiratory distress syndrome. Crit Care Med 1987;15:844–9.[ISI][Medline]
- Dive A, Miesse C, Galanti L, et al. Effect of erythromycin on gastric motility in mechanically ventilated critically ill patients: a double-blind, randomized, placebo-controlled study. Crit Care Med 1995;23:1356–62.[ISI][Medline]
- Jimenez P, Torres A, Roca J, et al. Arterial oxygenation does not predict the outcome of patients with acute respiratory failure needing mechanical ventilation. Eur Respir J 1994;7:730–5.[Abstract]
- Roupie E, Lepage E, Wysocki M, et al. Prevalence, etiologies and outcome of the acute respiratory distress syndrome among hypoxemic ventilated patients. Intensive Care Med 1999;25:920–9.[ISI][Medline]
- Bone RC, Maunder R, Slotman G, et al.; the Prostaglandin E1 Study Group. An early test of survival in patients with the adult respiratory distress syndrome. The PaO2/FlO2 ratio and its differential response to conventional therapy. Chest 1989;96:849–51.[Abstract/Free Full Text]
- Valta P, Uusaro A, Nunes S, et al. Acute respiratory distress syndrome: frequency, clinical course, and costs of care. Crit Care Med 1999;27:2367–74.[ISI][Medline]
- Flaatten H, Gjerde S, Guttormsen AB, et al. Outcome after acute respiratory failure is more dependent on dysfunction in other vital organs than on the severity of the respiratory failure. Crit Care 2003;7:R72–7.[ISI][Medline]
Top
Abstract
Introduction
