One-year Outcome after Severe Pulmonary Exacerbation in Adults with Cystic Fibrosis
http://www.100md.com
美国呼吸和危急护理医学 2005年第1期
Service de Pneumologie
Service de Reeanimation Meedicale
Service d'Informatique Meedicale et de Biostatistique, Groupe Hospitalier Cochin, Assistance Publique Hpitaux de Paris, Facultee Cochin Port-Royal, Universitee Paris 5, Paris, France
ABSTRACT
We retrospectively studied the outcomes of adult patients with cystic fibrosis (CF) hospitalized for severe pulmonary exacerbations (69 cases) between January 1997 and June 2001. Cases were treated either in the Pulmonary Department (n = 46) or in the intensive care unit (ICU) (n = 23) depending on severity. Noninvasive mechanical ventilation was used in 61% (14 of 23) and 33% (15 of 46) of cases treated in the ICU and the Pulmonary Department groups, respectively. Invasive ventilation was necessary in 4 of 23 cases treated in the ICU. The 1-year survival rate was 52% (12 of 23) and 91% (42 of 46) in the ICU and the Pulmonary Department groups, respectively. Lung transplantation was performed in two patients from the ICU group and in five patients from the Pulmonary Department group after hospital discharge. Factors predictive of death were prior colonization with Burkholderia cepacia and rapid decline in FEV1 before admission and severity of exacerbations (severity of hypoxemia and hypercapnia, simplified acute physiology score II and logistic organ dysfunction (LOD) scores, requirement of noninvasive mechanical ventilation, and hospitalization in the ICU) in the univariate analysis and were prior colonization with B. cepacia, the severity of hypoxemia at admission, and hospitalization in the ICU in the multivariate analysis. In 1-year survivors, pulmonary exacerbation did not affect the progression of the disease.
Key Words: cystic fibrosis intensive care unit nasal intermittent positive pressure ventilation pulmonary exacerbation respiratory failure
Cystic fibrosis (CF) is the most common lethal autosomal recessive disorder in the white population. Over 80% of patients with CF die because of respiratory failure (1). In adult patients with CF, severe pulmonary exacerbations are frequently observed during the end stage of this progressive disease or as an acute complication of rather stable disease. Severe exacerbations are often life threatening. Twenty years ago, the poor prognosis of patients with CF requiring intensive care for respiratory failure led to the suggestion that it was best to avoid mechanical ventilation and to restrict treatment to palliative care (2). However, improvements in disease management (3) and recent technical progress in ventilatory support have led most CF centers to treat severe pulmonary exacerbations more aggressively. A recent study among patients with CF hospitalized in an intensive care unit (ICU) for exacerbations with respiratory failure showed that intensive treatment is beneficial for patients with potentially reversible complications (such as hemoptysis or pneumothorax) or when a lung transplant could be performed (4). However, the authors found that patients with respiratory failure caused by CF exacerbation who did not undergo lung transplant had a poor prognosis.
Besides the immediate risk of death, pulmonary exacerbations may be associated with considerable physiologic deteriorations, some of which may irreversibly worsen pulmonary status. In patients with CF who survive severe acute pulmonary exacerbations, little is known about the influence of such exacerbations on the clinical course of the disease.
Identifying the factors that may predict a high risk of death or influence the clinical course of the disease will help to guide decisions concerning the choice between intensive treatment and emergency lung transplant in some patients with CF with severe pulmonary exacerbations. We have retrospectively studied all adult patients with CF who were hospitalized in the Pulmonary Department or the ICU of the adult CF center at Cochin Hospital (Paris, France) between January 1997 and June 2001 for severe pulmonary exacerbations. The aims of this study were to evaluate the 1-year survival in these patients and to identify factors that may influence death after severe pulmonary exacerbation. In patients who survived, we also examined the influence of the pulmonary exacerbations on the clinical course of the disease during the 12 months after hospitalization.
METHODS
Patient Selection
This study was performed on a cohort of 245 adult patients with CF followed in our CF center in the Pulmonary Department at Cochin Hospital (Faculty Cochin Port-Royal, University Paris 5, France). The main clinical characteristics of this cohort are summarized in Table 1. We used the admission records to identify all patients who were admitted to either the Pulmonary Department or the ICU between January 1, 1997, and June 30, 2001, with severe pulmonary exacerbations. When a patient was admitted several times within a year for repeated pulmonary exacerbations, the first admission was considered as the index exacerbation, whereas the others were recorded as an adverse event during the follow-up of the preceding episode. Patients admitted after lung transplantation were excluded from the study.
CF was diagnosed on the basis of at least one typical clinical manifestation (bronchiectasis and/or steatorrhea) and at least one biological criterion (5, 6), namely a sweat chloride concentration exceeding 60 mM (quantitative pilocarpine iontophoresis) and/or two pathologic mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene analyzed as described previously (7eC9).
The clinical characteristics of the patients were obtained from medical records. Data recorded included age, sex, age at diagnosis of CF, CFTR mutations, the presence of pancreatic insufficiency, diabetes mellitus, and chronic bacterial colonization with Pseudomonas aeruginosa or other organisms (defined as the presence of the organism for more than 6 months at a sputum culture density of 106 cfu/ml in at least three tests). We also recorded body mass index and results of pulmonary function tests obtained in stable conditions before exacerbation. Stable conditions were defined as the best clinical status observed either during a routine visit or after a course of antibiotics in patients requiring frequent treatment.
Definition and Management of Severe Pulmonary Exacerbations
Definition of severe pulmonary exacerbation and criteria for admission in the Pulmonary Department or the ICU.
Severe pulmonary exacerbation was defined as an event associated with deterioration of respiratory function requiring hospitalization. Criteria for hospital admission included deterioration of arterial blood gas characterized by a decrease in PaO2 of 10 mm Hg or more and/or an increase in PaCO2 of 5 mm Hg or more as compared with basal conditions associated with the occurrence of either pneumothorax or hemoptysis or with at least two or more of the following respiratory signs: increased sputum production, dyspnea, fever of more than 38°C for at least 48 hours, or new radiographic opacities.
Patients were referred to the ICU if they elicited at least one of the following criteria: respiratory acidosis with an arterial pH of less than 7.30, neurologic signs of hypercapnia, and clinical evidence of respiratory muscle fatigue (i.e., unstable clinical respiratory state requiring close monitoring). Other patients were hospitalized in the Pulmonary Department. Patients who refused to be hospitalized in the ICU were also treated in the Pulmonary Department and were analyzed as being part of the Pulmonary Department group. Patients who were first hospitalized in the Pulmonary Department before being transferred to the ICU for deterioration of their pulmonary status were analyzed in the ICU group. The severity of pulmonary exacerbation at admission was assessed from arterial blood gas, the new simplified acute physiology score, and logistic organ dysfunction (LOD) score values at admission (10, 11).
Infection was recognized as the most likely cause of exacerbation when it was associated with at least two or more of the following respiratory signs: increased sputum production, dyspnea, fever of more than 38°C for at least 48 hours, or new radiographic opacities and when no other identifiable cause such as pneumothorax or hemoptysis was found. Infection was also considered as the most likely cause of exacerbation in patients showing direct or indirect consequences of airway infection, such as an increase in mucus plugging, clinical evidence of right heart failure or muscle fatigue, when no other cause of exacerbation could be clearly identified.
Management of Pulmonary Exacerbation
Medical management of pulmonary exacerbation was done according to published guidelines (12). All patients had chest physiotherapy at least twice daily except for those who had hemoptysis or pneumothorax. Twenty percent of patients received regular treatment with RhDNase inhalation, and all patients received various bronchodilator medications during the exacerbation period. Every patient received intravenous antibiotics associating at least two antibiotics chosen according to the sensitivity of the bacteria isolated from sputum. Oxygen therapy was administered when PaO2 was less than 60 mm Hg and was adjusted to maintain if possible PaO2 of more than 60 mm Hg without a further increasing PaCO2. Ventilatory support was used for respiratory failure or to prevent respiratory failure when the patient's state worsened. The first step of ventilatory support consisted of noninvasive mechanical ventilation (NIV) using the pressure-support mode performed with conventional ventilators (Evita II Dura Drager or Servo 300 Siemens) in the ICU or bilevel positive pressure ventilators (Helia, Saime, S.A., Savigny-le-Temple, France) in the Pulmonary Department applied with a nasofacial mask or full facial mask. Ventilators were initially set to achieve a tidal volume of more than 5 ml/kg of body weight and a respiratory rate of less than 30 breaths/minute. The positive end-expiratory pressure was set at 5 cm of water, and the fraction of inspired oxygen was titrated to maintain arterial oxygen saturation above 90%. The ventilator settings were subsequently adjusted as necessary to obtain the best patient's comfort and to maintain adequate arterial blood gas and an arterial pH above 7.35. When respiratory failure occurred despite NIV, invasive positive pressure ventilation via an endotracheal tube was initiated. Enteral nutrition with adequate pancreatic enzyme and vitamin supplementation was given either orally or via a nasogastric catheter if necessary. Psychologic help was provided when necessary and accepted by the patient.
Hemoptysis was treated with intravenous Glypressin and bronchial arterial embolization if necessary. Pneumothorax was treated by chest tube aspiration.
Study Outcomes
The primary outcome was the 1-year mortality rate and survival after hospital admission. Clinical characteristics of the patients before severe exacerbation and the criteria of severity of exacerbations were used to examine factors that may have influenced survival. Outcomes of pulmonary exacerbations during the 1-year follow-up included also the durations of hospitalization and of ventilatory support, the frequency of new episodes of severe pulmonary exacerbations requiring hospital admissions, and the number of patients who received a lung transplant.
To determine whether pulmonary exacerbation influenced the course of the disease in patients who survived, we compared clinical data collected 1 year before the exacerbation and 1 year after hospital discharge. The following variables were analyzed: body mass index, number of intravenous antibiotic courses per year, the use of long-term oxygen treatment and/or NIV at home, FEV1 recorded in stable conditions. Using a linear regression analysis, we calculated and compared the slopes of FEV1 decrease 1 year before and 1 year after exacerbation in each survivor.
Statistical Analysis
When a patient was admitted several times for severe exacerbations at more than 12 months apart, exacerbations were analyzed as new and independent events. Therefore, data (including the 1-year mortality rate) were expressed as the percentage of total cases of exacerbations. Clinical and biological parameters were expressed in percentages or as means and SDs. Because of the nonnormal distribution of several variables and to small numbers of subjects in certain groups of interest, nonparametric statistical methods were used to examine relationships between variables when appropriate (Wilcoxon's tests, pairwise Wilcoxon's tests, Fisher's exact tests, and Mac Nemar tests). All p values are two tailed. The p values below 0.05 were considered to indicate significance.
Survival was analyzed by life-table analysis (Kaplan-Meier). Prognostic values of factors were investigated using Cox's proportional hazards models. These models were also used to study the effect of several factors simultaneously and to look for thresholds or cutoff values of predicators by testing second order polynomial effects. The final prognostic model was constructed in several stages. First, the model had to predict outcome from clinical characteristics of patients in a stable state before admission. In the second stage, the clinical characteristics at admission and care variables were added to significant independent predictors of the preceding stage (these variables were tested on an intention-to-treat basis). The last stage was to test interactions between the predictive variables of the final model. The results of analysis restricted to the first admission of the period for each patient were very similar to those of the entire group (all index admissions) and are therefore not shown. All calculations were performed using the SAS package (SAS Software Release 8.1; SAS Institutes, Cary, NC).
RESULTS
Clinical Characteristics of Patients
A total of 69 severe pulmonary exacerbations were identified, involving 57 patients. Twenty-three of these exacerbations (involving 19 patients) were treated in the ICU and 46 (involving 38 patients) in the Pulmonary Department. Two of the patients admitted to the Pulmonary Department fulfilled the criteria for admission to the ICU but refused to be hospitalized in the ICU. Nine of the 23 exacerbations admitted to the ICU were initially treated in the Pulmonary Department and then transferred to the ICU because of rapid deterioration of their pulmonary status. The clinical characteristics of the patients measured in stable conditions before admission were similar in the two groups (Table 2). The time intervals (mean ± SD) between recording of the stable values and admission for exacerbation were 2.2 ± 1.8 months and 1.8 ± 1.4 months for the ICU and the Pulmonary Department groups, respectively. Five patients (two in the ICU and three in the Pulmonary Department) were on a waiting list for lung transplant before admission.
Description of Severe Pulmonary Exacerbations
Infection was the most frequent cause of severe exacerbation in both the ICU group (12 of 23 or 52%) and the Pulmonary Department group (36 of 46 or 78%). We were able to compare bacterial pathogens isolated from the sputum before exacerbation (chronic colonization) and during exacerbation in 46 of 48 of these cases. New bacterial pathogens were found during exacerbation in two patients in the ICU group (one with Staphylococcus aureus and one with Stomatococcus mucilaginosus) and in five patients in the Pulmonary Department group (one patient with each of the following: S. aureus, Stenotrophomonas maltophilia, Alcaligenes xylosoxidans; one patient with both Alcaligenes xylosoxidans and S. aureus; and one with both Escherichia coli + H. influenzae). Other causes of exacerbation were hemoptysis and pneumothorax, both of which were more frequently observed in the group treated in the ICU group (30% and 17%, respectively) than in the Pulmonary Department group (17% and 4%, respectively). Clinical severity of pulmonary exacerbation at admission was greater in patients treated in the ICU than in patients treated in the Pulmonary Department (Table 3), as reflected by the simplified acute physiology score II and the LOD scores. Ventilatory support was necessary in most ICU cases. NIV was used in 61% of these cases (14 of 23), and invasive mechanical ventilation with an endotracheal tube was necessary in 17% (4 of 23) of cases because of failure of NIV. NIV was performed in 33% (15 of 46) of the cases treated in the Pulmonary Department. The durations of NIV and of hospitalization were similar in the ICU and the Pulmonary Department.
Survival and Predictors of Death
The 1-year survival rate was 78% in the total population. It was 52.2% (12 of 23) in cases treated in the ICU and 91.3% (42 of 46) in cases treated in the Pulmonary Department (p < 0.001) (Figure 1). Deaths were observed whatever the initial cause of exacerbation (Table 4). In the cases with exacerbation caused by infection, none of the seven cases for which new bacterial pathogens were isolated in sputum died. More than half of the deaths in the ICU group (6 of 11) occurred during hospitalization. The other five deaths occurred after hospital discharge between 1eC8 months after admission. In the ICU group, the four patients who required invasive mechanical ventilation using an endotracheal tube after failure of NIV died. Two of the four Pulmonary Department patients who died fulfilled the criteria for being hospitalized in the ICU but refused. None of the four patients died while in hospital. According to our definition of groups, the mortality rate in the group in the Pulmonary Department group did not include nine cases that were subsequently transferred to the ICU, six of whom died.
Univariate analysis showed that bronchial colonization with Burkholderia cepacia and rapid decline in FEV1 before admission were predictive of death after severe pulmonary exacerbation (Table 5). Among patients' characteristics at admission, predictors of death were the severity of hypoxemia in room air and of hypercapnia under oxygen, simplified acute physiology score II and LOD scores, and requirement for NIV (Table 6). Concerning the causes of exacerbation, neither pneumothorax nor hemoptysis influenced the risk of death. Multivariate analysis showed that independent factors influencing the risk of death were prior colonization with B. cepacia, hospitalization in the ICU, and the severity of hypoxemia at admission, both of which reflect the severity of exacerbations (Table 7). The slope of FEV1 decline before admission was steeper for patients colonized with B. cepacia than for patients colonized with other microorganisms (eC1.32 ± 1.14 and eC0.41 ± 1.20, p < 0.05, respectively). This association between colonization with B. cepacia and a decline in FEV1 before admission probably explains the loss of statistical significance of the rapid decline in FEV1 as predictor of death in the multivariate analysis.
Hospital Readmission and Lung Transplant after Severe Pulmonary Exacerbations within the 1-year Follow-up
During the year after exacerbation, hospital readmissions for severe pulmonary exacerbations occurred in 8 patients (34%; total of eight new exacerbations) from the ICU group and in 12 patients (26%; total of 15 new exacerbations) from the Pulmonary Department group. The 8 new exacerbations from the ICU group required hospitalization in the ICU, whereas the 15 new exacerbations from the Pulmonary Department group were managed in the Pulmonary Department. Seven of the 8 patients from the ICU group and 2 of the 12 patients from the Pulmonary Department group died during the 1-year follow-up.
During the 1 year after their episode of severe pulmonary exacerbation, seven patients received a lung transplant. Five of these patients were in the group treated in the Pulmonary Department (delay of 0.5 to 6 months after hospital discharge), and two were in the ICU group (delay of 1.5 and 8 months after hospital discharge). All of these patients were alive 1 year after exacerbation. Among the patients who benefited from lung transplant, only two of them (in the group treated in the Pulmonary Department) were on a waiting list for lung transplant before admission. Others were put on a transplant list either during or after exacerbation.
Influence of Severe Pulmonary Exacerbation on Progression of the Respiratory and Nutritional Status in 1-year Survivors
The pulmonary characteristics of 1-year survivors were compared before and after severe pulmonary exacerbation (Table 8). FEV1 values were significantly lower 1 year after admission. However, the rate of decline of FEV1 was similar during the year preceding and after pulmonary exacerbation. More patients required long-term oxygen treatment or chronic use of NIV after severe pulmonary exacerbation. However, exacerbations were not associated with reductions in body mass index or increases in the number of antibiotic courses per year.
DISCUSSION
The 1-year mortality rate of 22% observed in our total patient population with CF hospitalized for severe pulmonary exacerbation must take into account the differences in the results obtained in the ICU and the Pulmonary Department. In the group of patients in whom pulmonary exacerbation could be managed in the Pulmonary Department, the 1-year mortality rate was 8.7%, whereas it was 47.8% in the group of patients requiring treatment in the ICU. This difference would have been even greater if we had taken into account the fact that two of the four patients who died in the Pulmonary Department should have been hospitalized in the ICU but refused. The difference in the death rate between the two groups was not explained by a difference in the severity of the disease before exacerbation, as age, nutritional status, pulmonary function, nature of bacterial colonization, pancreatic status, and type of CFTR mutation were similar in the two subpopulations. The difference in the mortality rate was mostly explained by the severity of exacerbation, which was much greater in the ICU group than in the group of patients treated in the Pulmonary Department. This is supported by our finding that the severity of arterial blood gas disturbance at admission (PaO2 and PaCO2), the LOD and simplified acute physiology score II scores, and the use of ventilatory support during hospitalization, all of which reflect the severity of exacerbation, were factors predictive of death in the univariate analysis. The multivariate analysis confirmed these findings, showing that hospitalization in the ICU and severity of hypoxemia were predictive of death. It seems logical that the severity of exacerbation would explain deaths that occurred during hospitalization or shortly after exacerbations. In fact, none of the patients treated in the Pulmonary Department with less severe exacerbations died during hospitalization. In contrast, more than half of the deaths among patients hospitalized in the ICU (i.e., with more severe of exacerbation) occurred during hospitalization.
Severe pulmonary exacerbations affected mortality both during the period immediately after hospitalization and throughout the following year. Indeed, half of the deaths observed in our study occurred after hospital discharge and after the first 2 months after hospital admission. Severe pulmonary exacerbations appear to worsen dramatically the course of the disease, leading to relapse of severe pulmonary exacerbations and to death in some patients. Thus, in patients hospitalized in the ICU, one-third (8 of 23) of the cases were readmitted to the ICU for exacerbation during the year after the first episode; seven of these cases died. Similarly, one-fourth (12 of 46) of the Pulmonary Department cases were readmitted to the Pulmonary Department for severe exacerbation; two of these cases died. Clinical characteristics before admission that were found to influence the 1-year mortality rate in our study were prior colonization with B. cepacia and a rapid decline in FEV1. The former was found to be significant both in the univariate and multivariate analysis, whereas the latter was only significant in the univariate analysis. We found that patients colonized with B. cepacia also had a steeper decline in FEV1 values before admission. This association is probably responsible for the loss of significance of a rapid decline in FEV1 in the multivariate analysis. Our results are consistent with those of previous studies on the risk of mortality in the CF population outside of exacerbation. In a study of a cohort of patients with CF, a low FEV1 (less than 30% predicted) was associated with a poor prognosis, in particular in female subjects and younger patients (13). A rapid decline in FEV1 rather than a low FEV1 appeared to be the best predictor of death in patients with severely compromised lung function (14). The influence of colonization with B. cepacia on the risk of death has been clearly established in other studies, with a particularly poor prognosis for patients with B. cepacia from genomovar III (15, 16). The B. cepacia identified in the patients in our study were all from genomovar II.
In a previous study of 42 patients with CF hospitalized in an ICU for pulmonary exacerbations, the 1-year survival rate was 40% (17 of 42); most of the survivors had benefited from a lung transplant (14 of 17), whereas only three patients who were still alive at 1 year had not had a lung transplant (4). These results led the authors to conclude that treatment in an ICU can be successful when exacerbation is due to an acute reversible cause such as hemoptysis or pneumothorax or in patients for whom lung transplantation is an imminent option. The 1-year survival rate in our study was 52% (12 of 23) among patients requiring intensive care and 58% (7 of 12 patients) after excluding patients hospitalized for hemoptysis and pneumothorax. Neither hemoptysis nor pneumothorax, two potentially reversible causes of exacerbations, was associated with a lower risk of death in our study. Unlike the conclusions made by Sood and colleagues (4), our results suggest that the patients with CF with severe exacerbations should be treated in the ICUs even if they do not undergo lung transplant, as severe pulmonary exacerbations that require treatment in the ICU were associated with a 1-year survival over 40% (11 of 21 patients, after excluding the 2 patients who benefited from a lung transplantation). In our study, only two of the patients admitted in the ICU were on a waiting list for lung transplant prior to exacerbation; one of these patients died 1 month after hospital discharge before being transplanted, and the other one survived without lung transplant. The difference between our results and those of Sood and colleagues may be due to differences in the severity of exacerbation. This hypothesis cannot be evaluated, as Sood and colleagues did not provide information on the severity of exacerbation at admission. Ventilatory support was necessary in a similar proportion of exacerbations in the Sood and colleague's study (50 of 65 or 77%) and in our study (18 of 23 or 78%). However, one major difference between these two studies is the ventilatory support technique used; invasive mechanical ventilation with an endotracheal tube was used in 49% of exacerbations (32 of 65) in the study by Sood and coworkers and in only 17% (4 of 23) of our ICU cases. In contrast, NIV was used in 60% (14 of 23) of our ICU cases and just 28% (18 of 65) of the patients studied by Sood and associates. Although we cannot exclude the possibility that the higher rate of invasive ventilation with an endotracheal tube in the study by Sood and coworkers reflected greater severity, NIV may be associated with a better outcome than invasive ventilation in cases of severe CF exacerbation. This is supported by our findings that 50% of ICU exacerbations treated by NIV survived. The 1-year mortality rate is much higher when invasive mechanical ventilation is performed after failure of NIV, as shown by the fact that all four of the patients who required invasive mechanical ventilation in our study died. Similar results were reported in another series of 16 patients with CF treated in an ICU where the 8 (50%) survivors who were discharged were treated with NIV, whereas the other 8 patients who required invasive mechanical ventilation after failure of NIV died (17). Invasive mechanical ventilation using an endotracheal tube impairs coughing and secretion clearance and increases the risk of lethal pulmonary infection. The patients in the study by Sood and coworkers were recruited between 1990 and 1998, whereas our patients were enrolled study between 1997 and mid 2001. During the more recent period, techniques for ventilatory support using NIV have improved largely because of improved knowledge of the technique itself and improvements in the efficiency of the ventilators and facial masks. NIV has been successfully used in the management of stable patients with severe CF awaiting transplantation (18eC21). Therefore, our study suggests that in patients requiring respiratory support for acute pulmonary exacerbation NIV, when possible, is beneficial to patients with CF.
Our patients treated in the Pulmonary Department had a fairly good prognosis: no patients died during hospitalization and the 1-year survival rate was over 91%, even though two of these patients fulfilled the criteria for being treated in the ICU. Fifteen of these 46 exacerbations required NIV. This fairly good prognosis was observed despite the fact that after hospital discharge 26% of the Pulmonary Department cases had an unstable disease status requiring hospital readmission for severe pulmonary exacerbation with 1-year follow-up. Five of these patients underwent lung transplantation between 0.5 and 6 months after hospital discharge. However, it is important to remember that, by definition, this group of patients did not include nine patients initially treated in the Pulmonary Department who were subsequently transferred to the ICU because of a rapid deterioration of their pulmonary status despite NIV. These results indicate that in CF, pulmonary exacerbation without immediate need for ICU can be managed safely and successfully in a pulmonary department providing that the medical team is familiar with the treatment of patients with CF and with NIV and that the pulmonary department is closely linked to an ICU trained to treat patients with CF in case of worsening of the respiratory status.
The effects of pulmonary exacerbations on the course of the disease in patients who survive after exacerbation are unknown. A comparison of pulmonary function and nutritional status of survivors from both subgroups 1 year before and after exacerbation shows that the course of the disease was not modified by exacerbations. We found that the rates of decrease in FEV1 during the year preceding and after exacerbations were similar, indicating that the lower values of FEV1 observed 1 year after exacerbation reflect the natural degradation of pulmonary function caused by disease progression. The number of antibiotic courses and the nutritional status were also unchanged after exacerbation. However, more treatment was needed to maintain pulmonary function 1 year after exacerbation, as shown by the greater number of patients requiring long-term oxygen therapy (all but one patient) and the chronic use of NIV (50% of the patients). These findings reflect the continually worsening nature of the disease rather than consequences of exacerbations.
In conclusion, our results show that intensive treatment is beneficial for patients with CF with severe pulmonary exacerbation even if they do not undergo lung transplant. Less severe pulmonary exacerbations can be treated successfully in a pulmonary department with a good short- and long-term prognosis, providing that NIV can be performed and that a close collaboration exists with an ICU trained to treat patients with CF. Poor survival after exacerbation is directly related to the severity of exacerbation, which is reflected by the severity of hypoxemia at admission and the requirement for hospitalization in the ICU. Prior colonization with B. cepacia and a rapid decline in FEV1 were both shown to be associated with an increased risk of death in these patients. The presence of these factors in patients with severe exacerbations may indicate urgent transplant evaluation. In patients surviving severe exacerbations, the progression of the disease (pulmonary and nutritional status) is unaffected by pulmonary exacerbations.
REFERENCES
Davis PB. Clinical pathophysiology and manifestations of lung disease. In: Yankaskas JR, Knowles M, editors. Cystic fibrosis in adults, 1st ed. Philadelphia: Lippincott-Raven; 1999. pp. 45eC67.
di Sant'Agnese PA, Davius PB. Assisted ventilation for patients with cystic fibrosis. JAMA 1978;239:1851eC1854.
Yankaskas JR, Marshall BC, Sufian B, Simon RH, Rodman D. Cystic fibrosis adult care consensus report. Chest 2004;125:1SeC39S.
Sood N, Paradowski LJ, Yankaskas J. Outcomes of intensive care unit care in adults with cystic fibrosis. Am J Respir Crit Care Med 2001;163:335eC338.
Rosenstein BJ, Cutting JR. The diagnosis of CF: a consensus statement: Cystic Fibrosis Foundation Consensus Panel. J Pediatr 1998;132:589eC595.
Stern RC. The diagnosis of cystic fibrosis. N Engl J Med 1997;336:487eC491.
Bienvenu T, Cazeneuve C, Kaplan JC, Beldjord C. Mutation heterogeneity of cystic fibrosis in France: screening by denaturing gradient gel electrophoresis using psoralen-modified oligonucleotide. Hum Mutat 1995;6:23eC29.
Highsmith WE, Burch LH, Zhou Z, Olsen JC, Boat TE, Spock A, Gorvoy JD, Quittell L, Friedman J, Silverman LM, et al. A novel mutation in the cystic fibrosis gene in patients with pulmonary disease but normal sweat chloride concentrations. N Engl J Med 1994;33:974eC980.
Chillon M, Casals T, Mercier B, Bassas L, Lissens W, Silber S, Romey M-C, Ruiz-Romero J, Verlingue C, Claustres M, et al. Mutations in the cystic fibrosis gene in patients with congenital absence of the vas deferens. N Engl J Med 1995;332:1475eC1480.
Le Gall JR, Lemeshow S, Saulnier F. New simplified acute physiology score (SAPS II) based on European/North American multicenter study. JAMA 1993;270:2957eC2963.
Le Gall JR, Klar J, Lemeshow S, Saulnier F, Alberti C, Artigas A, Teres D. The logistic organ dysfunction system. A new way to assess organ dysfunction in intensive care unit: ICU scoring system. JAMA 1996;276:802eC810.
Schidlow D, Taussig L, Knowles M. Cystic Fibrosis Foundation Consensus Conference report on pulmonary complications of cystic fibrosis. Pediatr Pulmonol 1993;15:187eC198.
Kerem E, Reisman J, Corey M, Canny GJ, Levison H. Prediction of mortality in patients with cystic fibrosis. N Engl J Med 1992;326:1187eC1191.
Milla CE, Warwick WJ. Risk of death in cystic fibrosis patients with severely compromised lung function. Chest 1998;113:1230eC1234.
Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med 2003;168:918eC951.
Mudhi KM, O'Hickey S, Smith G, Stableforth DE. Rate of decline of lung function in patients with cystic fibrosis colonised by Pseudomonas cepacia. Thorax 1993;48:1088.
Cadiergue V, Philit F, Langevin B, Durieu I, Bertocchi M, Gueerin C, Robert D. Devenir des adultes mucoviscidosiques admis en reeanimation pour deecompensation respiratoire. Rev Mal Respir 2002;19:425eC430.
Hodson ME, Madden BP, Steven MH, Tsang VT, Yacoub MH. Non-invasive mechanical ventilation for cystic fibrosis patients, a potential bridge to transplantation. Eur Respir J 1991;4:524eC527.
Granton JT, Kesten S. The acute effects of nasal positive pressure ventilation in patients with advanced cystic fibrosis. Chest 1998;113:1013eC1018.
Hill AT, Edenborough FP, Cayton RM, Stableforth DE. Long-term nasal intermittent positive pressure ventilation in patients with cystic fibrosis and hypercapnic respiratory failure (1991eC1996). Respir Med 1998;92:523eC526.
Piper AJ, Parker S, Torzillo PJ, Sullivan CE, Bye PT. Nocturnal nasal IPPV stabilizes patients with cystic fibrosis and hypercapnic respiratory failure. Chest 1992;102:846eC850.(Madiha Ellaffi, Christoph)
Service de Reeanimation Meedicale
Service d'Informatique Meedicale et de Biostatistique, Groupe Hospitalier Cochin, Assistance Publique Hpitaux de Paris, Facultee Cochin Port-Royal, Universitee Paris 5, Paris, France
ABSTRACT
We retrospectively studied the outcomes of adult patients with cystic fibrosis (CF) hospitalized for severe pulmonary exacerbations (69 cases) between January 1997 and June 2001. Cases were treated either in the Pulmonary Department (n = 46) or in the intensive care unit (ICU) (n = 23) depending on severity. Noninvasive mechanical ventilation was used in 61% (14 of 23) and 33% (15 of 46) of cases treated in the ICU and the Pulmonary Department groups, respectively. Invasive ventilation was necessary in 4 of 23 cases treated in the ICU. The 1-year survival rate was 52% (12 of 23) and 91% (42 of 46) in the ICU and the Pulmonary Department groups, respectively. Lung transplantation was performed in two patients from the ICU group and in five patients from the Pulmonary Department group after hospital discharge. Factors predictive of death were prior colonization with Burkholderia cepacia and rapid decline in FEV1 before admission and severity of exacerbations (severity of hypoxemia and hypercapnia, simplified acute physiology score II and logistic organ dysfunction (LOD) scores, requirement of noninvasive mechanical ventilation, and hospitalization in the ICU) in the univariate analysis and were prior colonization with B. cepacia, the severity of hypoxemia at admission, and hospitalization in the ICU in the multivariate analysis. In 1-year survivors, pulmonary exacerbation did not affect the progression of the disease.
Key Words: cystic fibrosis intensive care unit nasal intermittent positive pressure ventilation pulmonary exacerbation respiratory failure
Cystic fibrosis (CF) is the most common lethal autosomal recessive disorder in the white population. Over 80% of patients with CF die because of respiratory failure (1). In adult patients with CF, severe pulmonary exacerbations are frequently observed during the end stage of this progressive disease or as an acute complication of rather stable disease. Severe exacerbations are often life threatening. Twenty years ago, the poor prognosis of patients with CF requiring intensive care for respiratory failure led to the suggestion that it was best to avoid mechanical ventilation and to restrict treatment to palliative care (2). However, improvements in disease management (3) and recent technical progress in ventilatory support have led most CF centers to treat severe pulmonary exacerbations more aggressively. A recent study among patients with CF hospitalized in an intensive care unit (ICU) for exacerbations with respiratory failure showed that intensive treatment is beneficial for patients with potentially reversible complications (such as hemoptysis or pneumothorax) or when a lung transplant could be performed (4). However, the authors found that patients with respiratory failure caused by CF exacerbation who did not undergo lung transplant had a poor prognosis.
Besides the immediate risk of death, pulmonary exacerbations may be associated with considerable physiologic deteriorations, some of which may irreversibly worsen pulmonary status. In patients with CF who survive severe acute pulmonary exacerbations, little is known about the influence of such exacerbations on the clinical course of the disease.
Identifying the factors that may predict a high risk of death or influence the clinical course of the disease will help to guide decisions concerning the choice between intensive treatment and emergency lung transplant in some patients with CF with severe pulmonary exacerbations. We have retrospectively studied all adult patients with CF who were hospitalized in the Pulmonary Department or the ICU of the adult CF center at Cochin Hospital (Paris, France) between January 1997 and June 2001 for severe pulmonary exacerbations. The aims of this study were to evaluate the 1-year survival in these patients and to identify factors that may influence death after severe pulmonary exacerbation. In patients who survived, we also examined the influence of the pulmonary exacerbations on the clinical course of the disease during the 12 months after hospitalization.
METHODS
Patient Selection
This study was performed on a cohort of 245 adult patients with CF followed in our CF center in the Pulmonary Department at Cochin Hospital (Faculty Cochin Port-Royal, University Paris 5, France). The main clinical characteristics of this cohort are summarized in Table 1. We used the admission records to identify all patients who were admitted to either the Pulmonary Department or the ICU between January 1, 1997, and June 30, 2001, with severe pulmonary exacerbations. When a patient was admitted several times within a year for repeated pulmonary exacerbations, the first admission was considered as the index exacerbation, whereas the others were recorded as an adverse event during the follow-up of the preceding episode. Patients admitted after lung transplantation were excluded from the study.
CF was diagnosed on the basis of at least one typical clinical manifestation (bronchiectasis and/or steatorrhea) and at least one biological criterion (5, 6), namely a sweat chloride concentration exceeding 60 mM (quantitative pilocarpine iontophoresis) and/or two pathologic mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene analyzed as described previously (7eC9).
The clinical characteristics of the patients were obtained from medical records. Data recorded included age, sex, age at diagnosis of CF, CFTR mutations, the presence of pancreatic insufficiency, diabetes mellitus, and chronic bacterial colonization with Pseudomonas aeruginosa or other organisms (defined as the presence of the organism for more than 6 months at a sputum culture density of 106 cfu/ml in at least three tests). We also recorded body mass index and results of pulmonary function tests obtained in stable conditions before exacerbation. Stable conditions were defined as the best clinical status observed either during a routine visit or after a course of antibiotics in patients requiring frequent treatment.
Definition and Management of Severe Pulmonary Exacerbations
Definition of severe pulmonary exacerbation and criteria for admission in the Pulmonary Department or the ICU.
Severe pulmonary exacerbation was defined as an event associated with deterioration of respiratory function requiring hospitalization. Criteria for hospital admission included deterioration of arterial blood gas characterized by a decrease in PaO2 of 10 mm Hg or more and/or an increase in PaCO2 of 5 mm Hg or more as compared with basal conditions associated with the occurrence of either pneumothorax or hemoptysis or with at least two or more of the following respiratory signs: increased sputum production, dyspnea, fever of more than 38°C for at least 48 hours, or new radiographic opacities.
Patients were referred to the ICU if they elicited at least one of the following criteria: respiratory acidosis with an arterial pH of less than 7.30, neurologic signs of hypercapnia, and clinical evidence of respiratory muscle fatigue (i.e., unstable clinical respiratory state requiring close monitoring). Other patients were hospitalized in the Pulmonary Department. Patients who refused to be hospitalized in the ICU were also treated in the Pulmonary Department and were analyzed as being part of the Pulmonary Department group. Patients who were first hospitalized in the Pulmonary Department before being transferred to the ICU for deterioration of their pulmonary status were analyzed in the ICU group. The severity of pulmonary exacerbation at admission was assessed from arterial blood gas, the new simplified acute physiology score, and logistic organ dysfunction (LOD) score values at admission (10, 11).
Infection was recognized as the most likely cause of exacerbation when it was associated with at least two or more of the following respiratory signs: increased sputum production, dyspnea, fever of more than 38°C for at least 48 hours, or new radiographic opacities and when no other identifiable cause such as pneumothorax or hemoptysis was found. Infection was also considered as the most likely cause of exacerbation in patients showing direct or indirect consequences of airway infection, such as an increase in mucus plugging, clinical evidence of right heart failure or muscle fatigue, when no other cause of exacerbation could be clearly identified.
Management of Pulmonary Exacerbation
Medical management of pulmonary exacerbation was done according to published guidelines (12). All patients had chest physiotherapy at least twice daily except for those who had hemoptysis or pneumothorax. Twenty percent of patients received regular treatment with RhDNase inhalation, and all patients received various bronchodilator medications during the exacerbation period. Every patient received intravenous antibiotics associating at least two antibiotics chosen according to the sensitivity of the bacteria isolated from sputum. Oxygen therapy was administered when PaO2 was less than 60 mm Hg and was adjusted to maintain if possible PaO2 of more than 60 mm Hg without a further increasing PaCO2. Ventilatory support was used for respiratory failure or to prevent respiratory failure when the patient's state worsened. The first step of ventilatory support consisted of noninvasive mechanical ventilation (NIV) using the pressure-support mode performed with conventional ventilators (Evita II Dura Drager or Servo 300 Siemens) in the ICU or bilevel positive pressure ventilators (Helia, Saime, S.A., Savigny-le-Temple, France) in the Pulmonary Department applied with a nasofacial mask or full facial mask. Ventilators were initially set to achieve a tidal volume of more than 5 ml/kg of body weight and a respiratory rate of less than 30 breaths/minute. The positive end-expiratory pressure was set at 5 cm of water, and the fraction of inspired oxygen was titrated to maintain arterial oxygen saturation above 90%. The ventilator settings were subsequently adjusted as necessary to obtain the best patient's comfort and to maintain adequate arterial blood gas and an arterial pH above 7.35. When respiratory failure occurred despite NIV, invasive positive pressure ventilation via an endotracheal tube was initiated. Enteral nutrition with adequate pancreatic enzyme and vitamin supplementation was given either orally or via a nasogastric catheter if necessary. Psychologic help was provided when necessary and accepted by the patient.
Hemoptysis was treated with intravenous Glypressin and bronchial arterial embolization if necessary. Pneumothorax was treated by chest tube aspiration.
Study Outcomes
The primary outcome was the 1-year mortality rate and survival after hospital admission. Clinical characteristics of the patients before severe exacerbation and the criteria of severity of exacerbations were used to examine factors that may have influenced survival. Outcomes of pulmonary exacerbations during the 1-year follow-up included also the durations of hospitalization and of ventilatory support, the frequency of new episodes of severe pulmonary exacerbations requiring hospital admissions, and the number of patients who received a lung transplant.
To determine whether pulmonary exacerbation influenced the course of the disease in patients who survived, we compared clinical data collected 1 year before the exacerbation and 1 year after hospital discharge. The following variables were analyzed: body mass index, number of intravenous antibiotic courses per year, the use of long-term oxygen treatment and/or NIV at home, FEV1 recorded in stable conditions. Using a linear regression analysis, we calculated and compared the slopes of FEV1 decrease 1 year before and 1 year after exacerbation in each survivor.
Statistical Analysis
When a patient was admitted several times for severe exacerbations at more than 12 months apart, exacerbations were analyzed as new and independent events. Therefore, data (including the 1-year mortality rate) were expressed as the percentage of total cases of exacerbations. Clinical and biological parameters were expressed in percentages or as means and SDs. Because of the nonnormal distribution of several variables and to small numbers of subjects in certain groups of interest, nonparametric statistical methods were used to examine relationships between variables when appropriate (Wilcoxon's tests, pairwise Wilcoxon's tests, Fisher's exact tests, and Mac Nemar tests). All p values are two tailed. The p values below 0.05 were considered to indicate significance.
Survival was analyzed by life-table analysis (Kaplan-Meier). Prognostic values of factors were investigated using Cox's proportional hazards models. These models were also used to study the effect of several factors simultaneously and to look for thresholds or cutoff values of predicators by testing second order polynomial effects. The final prognostic model was constructed in several stages. First, the model had to predict outcome from clinical characteristics of patients in a stable state before admission. In the second stage, the clinical characteristics at admission and care variables were added to significant independent predictors of the preceding stage (these variables were tested on an intention-to-treat basis). The last stage was to test interactions between the predictive variables of the final model. The results of analysis restricted to the first admission of the period for each patient were very similar to those of the entire group (all index admissions) and are therefore not shown. All calculations were performed using the SAS package (SAS Software Release 8.1; SAS Institutes, Cary, NC).
RESULTS
Clinical Characteristics of Patients
A total of 69 severe pulmonary exacerbations were identified, involving 57 patients. Twenty-three of these exacerbations (involving 19 patients) were treated in the ICU and 46 (involving 38 patients) in the Pulmonary Department. Two of the patients admitted to the Pulmonary Department fulfilled the criteria for admission to the ICU but refused to be hospitalized in the ICU. Nine of the 23 exacerbations admitted to the ICU were initially treated in the Pulmonary Department and then transferred to the ICU because of rapid deterioration of their pulmonary status. The clinical characteristics of the patients measured in stable conditions before admission were similar in the two groups (Table 2). The time intervals (mean ± SD) between recording of the stable values and admission for exacerbation were 2.2 ± 1.8 months and 1.8 ± 1.4 months for the ICU and the Pulmonary Department groups, respectively. Five patients (two in the ICU and three in the Pulmonary Department) were on a waiting list for lung transplant before admission.
Description of Severe Pulmonary Exacerbations
Infection was the most frequent cause of severe exacerbation in both the ICU group (12 of 23 or 52%) and the Pulmonary Department group (36 of 46 or 78%). We were able to compare bacterial pathogens isolated from the sputum before exacerbation (chronic colonization) and during exacerbation in 46 of 48 of these cases. New bacterial pathogens were found during exacerbation in two patients in the ICU group (one with Staphylococcus aureus and one with Stomatococcus mucilaginosus) and in five patients in the Pulmonary Department group (one patient with each of the following: S. aureus, Stenotrophomonas maltophilia, Alcaligenes xylosoxidans; one patient with both Alcaligenes xylosoxidans and S. aureus; and one with both Escherichia coli + H. influenzae). Other causes of exacerbation were hemoptysis and pneumothorax, both of which were more frequently observed in the group treated in the ICU group (30% and 17%, respectively) than in the Pulmonary Department group (17% and 4%, respectively). Clinical severity of pulmonary exacerbation at admission was greater in patients treated in the ICU than in patients treated in the Pulmonary Department (Table 3), as reflected by the simplified acute physiology score II and the LOD scores. Ventilatory support was necessary in most ICU cases. NIV was used in 61% of these cases (14 of 23), and invasive mechanical ventilation with an endotracheal tube was necessary in 17% (4 of 23) of cases because of failure of NIV. NIV was performed in 33% (15 of 46) of the cases treated in the Pulmonary Department. The durations of NIV and of hospitalization were similar in the ICU and the Pulmonary Department.
Survival and Predictors of Death
The 1-year survival rate was 78% in the total population. It was 52.2% (12 of 23) in cases treated in the ICU and 91.3% (42 of 46) in cases treated in the Pulmonary Department (p < 0.001) (Figure 1). Deaths were observed whatever the initial cause of exacerbation (Table 4). In the cases with exacerbation caused by infection, none of the seven cases for which new bacterial pathogens were isolated in sputum died. More than half of the deaths in the ICU group (6 of 11) occurred during hospitalization. The other five deaths occurred after hospital discharge between 1eC8 months after admission. In the ICU group, the four patients who required invasive mechanical ventilation using an endotracheal tube after failure of NIV died. Two of the four Pulmonary Department patients who died fulfilled the criteria for being hospitalized in the ICU but refused. None of the four patients died while in hospital. According to our definition of groups, the mortality rate in the group in the Pulmonary Department group did not include nine cases that were subsequently transferred to the ICU, six of whom died.
Univariate analysis showed that bronchial colonization with Burkholderia cepacia and rapid decline in FEV1 before admission were predictive of death after severe pulmonary exacerbation (Table 5). Among patients' characteristics at admission, predictors of death were the severity of hypoxemia in room air and of hypercapnia under oxygen, simplified acute physiology score II and LOD scores, and requirement for NIV (Table 6). Concerning the causes of exacerbation, neither pneumothorax nor hemoptysis influenced the risk of death. Multivariate analysis showed that independent factors influencing the risk of death were prior colonization with B. cepacia, hospitalization in the ICU, and the severity of hypoxemia at admission, both of which reflect the severity of exacerbations (Table 7). The slope of FEV1 decline before admission was steeper for patients colonized with B. cepacia than for patients colonized with other microorganisms (eC1.32 ± 1.14 and eC0.41 ± 1.20, p < 0.05, respectively). This association between colonization with B. cepacia and a decline in FEV1 before admission probably explains the loss of statistical significance of the rapid decline in FEV1 as predictor of death in the multivariate analysis.
Hospital Readmission and Lung Transplant after Severe Pulmonary Exacerbations within the 1-year Follow-up
During the year after exacerbation, hospital readmissions for severe pulmonary exacerbations occurred in 8 patients (34%; total of eight new exacerbations) from the ICU group and in 12 patients (26%; total of 15 new exacerbations) from the Pulmonary Department group. The 8 new exacerbations from the ICU group required hospitalization in the ICU, whereas the 15 new exacerbations from the Pulmonary Department group were managed in the Pulmonary Department. Seven of the 8 patients from the ICU group and 2 of the 12 patients from the Pulmonary Department group died during the 1-year follow-up.
During the 1 year after their episode of severe pulmonary exacerbation, seven patients received a lung transplant. Five of these patients were in the group treated in the Pulmonary Department (delay of 0.5 to 6 months after hospital discharge), and two were in the ICU group (delay of 1.5 and 8 months after hospital discharge). All of these patients were alive 1 year after exacerbation. Among the patients who benefited from lung transplant, only two of them (in the group treated in the Pulmonary Department) were on a waiting list for lung transplant before admission. Others were put on a transplant list either during or after exacerbation.
Influence of Severe Pulmonary Exacerbation on Progression of the Respiratory and Nutritional Status in 1-year Survivors
The pulmonary characteristics of 1-year survivors were compared before and after severe pulmonary exacerbation (Table 8). FEV1 values were significantly lower 1 year after admission. However, the rate of decline of FEV1 was similar during the year preceding and after pulmonary exacerbation. More patients required long-term oxygen treatment or chronic use of NIV after severe pulmonary exacerbation. However, exacerbations were not associated with reductions in body mass index or increases in the number of antibiotic courses per year.
DISCUSSION
The 1-year mortality rate of 22% observed in our total patient population with CF hospitalized for severe pulmonary exacerbation must take into account the differences in the results obtained in the ICU and the Pulmonary Department. In the group of patients in whom pulmonary exacerbation could be managed in the Pulmonary Department, the 1-year mortality rate was 8.7%, whereas it was 47.8% in the group of patients requiring treatment in the ICU. This difference would have been even greater if we had taken into account the fact that two of the four patients who died in the Pulmonary Department should have been hospitalized in the ICU but refused. The difference in the death rate between the two groups was not explained by a difference in the severity of the disease before exacerbation, as age, nutritional status, pulmonary function, nature of bacterial colonization, pancreatic status, and type of CFTR mutation were similar in the two subpopulations. The difference in the mortality rate was mostly explained by the severity of exacerbation, which was much greater in the ICU group than in the group of patients treated in the Pulmonary Department. This is supported by our finding that the severity of arterial blood gas disturbance at admission (PaO2 and PaCO2), the LOD and simplified acute physiology score II scores, and the use of ventilatory support during hospitalization, all of which reflect the severity of exacerbation, were factors predictive of death in the univariate analysis. The multivariate analysis confirmed these findings, showing that hospitalization in the ICU and severity of hypoxemia were predictive of death. It seems logical that the severity of exacerbation would explain deaths that occurred during hospitalization or shortly after exacerbations. In fact, none of the patients treated in the Pulmonary Department with less severe exacerbations died during hospitalization. In contrast, more than half of the deaths among patients hospitalized in the ICU (i.e., with more severe of exacerbation) occurred during hospitalization.
Severe pulmonary exacerbations affected mortality both during the period immediately after hospitalization and throughout the following year. Indeed, half of the deaths observed in our study occurred after hospital discharge and after the first 2 months after hospital admission. Severe pulmonary exacerbations appear to worsen dramatically the course of the disease, leading to relapse of severe pulmonary exacerbations and to death in some patients. Thus, in patients hospitalized in the ICU, one-third (8 of 23) of the cases were readmitted to the ICU for exacerbation during the year after the first episode; seven of these cases died. Similarly, one-fourth (12 of 46) of the Pulmonary Department cases were readmitted to the Pulmonary Department for severe exacerbation; two of these cases died. Clinical characteristics before admission that were found to influence the 1-year mortality rate in our study were prior colonization with B. cepacia and a rapid decline in FEV1. The former was found to be significant both in the univariate and multivariate analysis, whereas the latter was only significant in the univariate analysis. We found that patients colonized with B. cepacia also had a steeper decline in FEV1 values before admission. This association is probably responsible for the loss of significance of a rapid decline in FEV1 in the multivariate analysis. Our results are consistent with those of previous studies on the risk of mortality in the CF population outside of exacerbation. In a study of a cohort of patients with CF, a low FEV1 (less than 30% predicted) was associated with a poor prognosis, in particular in female subjects and younger patients (13). A rapid decline in FEV1 rather than a low FEV1 appeared to be the best predictor of death in patients with severely compromised lung function (14). The influence of colonization with B. cepacia on the risk of death has been clearly established in other studies, with a particularly poor prognosis for patients with B. cepacia from genomovar III (15, 16). The B. cepacia identified in the patients in our study were all from genomovar II.
In a previous study of 42 patients with CF hospitalized in an ICU for pulmonary exacerbations, the 1-year survival rate was 40% (17 of 42); most of the survivors had benefited from a lung transplant (14 of 17), whereas only three patients who were still alive at 1 year had not had a lung transplant (4). These results led the authors to conclude that treatment in an ICU can be successful when exacerbation is due to an acute reversible cause such as hemoptysis or pneumothorax or in patients for whom lung transplantation is an imminent option. The 1-year survival rate in our study was 52% (12 of 23) among patients requiring intensive care and 58% (7 of 12 patients) after excluding patients hospitalized for hemoptysis and pneumothorax. Neither hemoptysis nor pneumothorax, two potentially reversible causes of exacerbations, was associated with a lower risk of death in our study. Unlike the conclusions made by Sood and colleagues (4), our results suggest that the patients with CF with severe exacerbations should be treated in the ICUs even if they do not undergo lung transplant, as severe pulmonary exacerbations that require treatment in the ICU were associated with a 1-year survival over 40% (11 of 21 patients, after excluding the 2 patients who benefited from a lung transplantation). In our study, only two of the patients admitted in the ICU were on a waiting list for lung transplant prior to exacerbation; one of these patients died 1 month after hospital discharge before being transplanted, and the other one survived without lung transplant. The difference between our results and those of Sood and colleagues may be due to differences in the severity of exacerbation. This hypothesis cannot be evaluated, as Sood and colleagues did not provide information on the severity of exacerbation at admission. Ventilatory support was necessary in a similar proportion of exacerbations in the Sood and colleague's study (50 of 65 or 77%) and in our study (18 of 23 or 78%). However, one major difference between these two studies is the ventilatory support technique used; invasive mechanical ventilation with an endotracheal tube was used in 49% of exacerbations (32 of 65) in the study by Sood and coworkers and in only 17% (4 of 23) of our ICU cases. In contrast, NIV was used in 60% (14 of 23) of our ICU cases and just 28% (18 of 65) of the patients studied by Sood and associates. Although we cannot exclude the possibility that the higher rate of invasive ventilation with an endotracheal tube in the study by Sood and coworkers reflected greater severity, NIV may be associated with a better outcome than invasive ventilation in cases of severe CF exacerbation. This is supported by our findings that 50% of ICU exacerbations treated by NIV survived. The 1-year mortality rate is much higher when invasive mechanical ventilation is performed after failure of NIV, as shown by the fact that all four of the patients who required invasive mechanical ventilation in our study died. Similar results were reported in another series of 16 patients with CF treated in an ICU where the 8 (50%) survivors who were discharged were treated with NIV, whereas the other 8 patients who required invasive mechanical ventilation after failure of NIV died (17). Invasive mechanical ventilation using an endotracheal tube impairs coughing and secretion clearance and increases the risk of lethal pulmonary infection. The patients in the study by Sood and coworkers were recruited between 1990 and 1998, whereas our patients were enrolled study between 1997 and mid 2001. During the more recent period, techniques for ventilatory support using NIV have improved largely because of improved knowledge of the technique itself and improvements in the efficiency of the ventilators and facial masks. NIV has been successfully used in the management of stable patients with severe CF awaiting transplantation (18eC21). Therefore, our study suggests that in patients requiring respiratory support for acute pulmonary exacerbation NIV, when possible, is beneficial to patients with CF.
Our patients treated in the Pulmonary Department had a fairly good prognosis: no patients died during hospitalization and the 1-year survival rate was over 91%, even though two of these patients fulfilled the criteria for being treated in the ICU. Fifteen of these 46 exacerbations required NIV. This fairly good prognosis was observed despite the fact that after hospital discharge 26% of the Pulmonary Department cases had an unstable disease status requiring hospital readmission for severe pulmonary exacerbation with 1-year follow-up. Five of these patients underwent lung transplantation between 0.5 and 6 months after hospital discharge. However, it is important to remember that, by definition, this group of patients did not include nine patients initially treated in the Pulmonary Department who were subsequently transferred to the ICU because of a rapid deterioration of their pulmonary status despite NIV. These results indicate that in CF, pulmonary exacerbation without immediate need for ICU can be managed safely and successfully in a pulmonary department providing that the medical team is familiar with the treatment of patients with CF and with NIV and that the pulmonary department is closely linked to an ICU trained to treat patients with CF in case of worsening of the respiratory status.
The effects of pulmonary exacerbations on the course of the disease in patients who survive after exacerbation are unknown. A comparison of pulmonary function and nutritional status of survivors from both subgroups 1 year before and after exacerbation shows that the course of the disease was not modified by exacerbations. We found that the rates of decrease in FEV1 during the year preceding and after exacerbations were similar, indicating that the lower values of FEV1 observed 1 year after exacerbation reflect the natural degradation of pulmonary function caused by disease progression. The number of antibiotic courses and the nutritional status were also unchanged after exacerbation. However, more treatment was needed to maintain pulmonary function 1 year after exacerbation, as shown by the greater number of patients requiring long-term oxygen therapy (all but one patient) and the chronic use of NIV (50% of the patients). These findings reflect the continually worsening nature of the disease rather than consequences of exacerbations.
In conclusion, our results show that intensive treatment is beneficial for patients with CF with severe pulmonary exacerbation even if they do not undergo lung transplant. Less severe pulmonary exacerbations can be treated successfully in a pulmonary department with a good short- and long-term prognosis, providing that NIV can be performed and that a close collaboration exists with an ICU trained to treat patients with CF. Poor survival after exacerbation is directly related to the severity of exacerbation, which is reflected by the severity of hypoxemia at admission and the requirement for hospitalization in the ICU. Prior colonization with B. cepacia and a rapid decline in FEV1 were both shown to be associated with an increased risk of death in these patients. The presence of these factors in patients with severe exacerbations may indicate urgent transplant evaluation. In patients surviving severe exacerbations, the progression of the disease (pulmonary and nutritional status) is unaffected by pulmonary exacerbations.
REFERENCES
Davis PB. Clinical pathophysiology and manifestations of lung disease. In: Yankaskas JR, Knowles M, editors. Cystic fibrosis in adults, 1st ed. Philadelphia: Lippincott-Raven; 1999. pp. 45eC67.
di Sant'Agnese PA, Davius PB. Assisted ventilation for patients with cystic fibrosis. JAMA 1978;239:1851eC1854.
Yankaskas JR, Marshall BC, Sufian B, Simon RH, Rodman D. Cystic fibrosis adult care consensus report. Chest 2004;125:1SeC39S.
Sood N, Paradowski LJ, Yankaskas J. Outcomes of intensive care unit care in adults with cystic fibrosis. Am J Respir Crit Care Med 2001;163:335eC338.
Rosenstein BJ, Cutting JR. The diagnosis of CF: a consensus statement: Cystic Fibrosis Foundation Consensus Panel. J Pediatr 1998;132:589eC595.
Stern RC. The diagnosis of cystic fibrosis. N Engl J Med 1997;336:487eC491.
Bienvenu T, Cazeneuve C, Kaplan JC, Beldjord C. Mutation heterogeneity of cystic fibrosis in France: screening by denaturing gradient gel electrophoresis using psoralen-modified oligonucleotide. Hum Mutat 1995;6:23eC29.
Highsmith WE, Burch LH, Zhou Z, Olsen JC, Boat TE, Spock A, Gorvoy JD, Quittell L, Friedman J, Silverman LM, et al. A novel mutation in the cystic fibrosis gene in patients with pulmonary disease but normal sweat chloride concentrations. N Engl J Med 1994;33:974eC980.
Chillon M, Casals T, Mercier B, Bassas L, Lissens W, Silber S, Romey M-C, Ruiz-Romero J, Verlingue C, Claustres M, et al. Mutations in the cystic fibrosis gene in patients with congenital absence of the vas deferens. N Engl J Med 1995;332:1475eC1480.
Le Gall JR, Lemeshow S, Saulnier F. New simplified acute physiology score (SAPS II) based on European/North American multicenter study. JAMA 1993;270:2957eC2963.
Le Gall JR, Klar J, Lemeshow S, Saulnier F, Alberti C, Artigas A, Teres D. The logistic organ dysfunction system. A new way to assess organ dysfunction in intensive care unit: ICU scoring system. JAMA 1996;276:802eC810.
Schidlow D, Taussig L, Knowles M. Cystic Fibrosis Foundation Consensus Conference report on pulmonary complications of cystic fibrosis. Pediatr Pulmonol 1993;15:187eC198.
Kerem E, Reisman J, Corey M, Canny GJ, Levison H. Prediction of mortality in patients with cystic fibrosis. N Engl J Med 1992;326:1187eC1191.
Milla CE, Warwick WJ. Risk of death in cystic fibrosis patients with severely compromised lung function. Chest 1998;113:1230eC1234.
Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med 2003;168:918eC951.
Mudhi KM, O'Hickey S, Smith G, Stableforth DE. Rate of decline of lung function in patients with cystic fibrosis colonised by Pseudomonas cepacia. Thorax 1993;48:1088.
Cadiergue V, Philit F, Langevin B, Durieu I, Bertocchi M, Gueerin C, Robert D. Devenir des adultes mucoviscidosiques admis en reeanimation pour deecompensation respiratoire. Rev Mal Respir 2002;19:425eC430.
Hodson ME, Madden BP, Steven MH, Tsang VT, Yacoub MH. Non-invasive mechanical ventilation for cystic fibrosis patients, a potential bridge to transplantation. Eur Respir J 1991;4:524eC527.
Granton JT, Kesten S. The acute effects of nasal positive pressure ventilation in patients with advanced cystic fibrosis. Chest 1998;113:1013eC1018.
Hill AT, Edenborough FP, Cayton RM, Stableforth DE. Long-term nasal intermittent positive pressure ventilation in patients with cystic fibrosis and hypercapnic respiratory failure (1991eC1996). Respir Med 1998;92:523eC526.
Piper AJ, Parker S, Torzillo PJ, Sullivan CE, Bye PT. Nocturnal nasal IPPV stabilizes patients with cystic fibrosis and hypercapnic respiratory failure. Chest 1992;102:846eC850.(Madiha Ellaffi, Christoph)