To our knowledge, this is the first study to investigate long term outcomes of children with CF admitted to PICU for respiratory failure. Two main conclusions can be drawn from this study: (1) Admission in PICU for acute respiratory failure is associated with poor outcomes, with only 35% of children being discharged alive without lung transplantation; (2) Referral to a lung transplant center must be considered for those surviving to the PICU stay as 36 months after discharge only 30% were still alive without lung transplantation and 60% deceased.
Mortality among our population of CF children with respiratory failure hospitalized in PICU was 17%. This is significantly higher than the number of 6.6% reported by Smith et al (17). However this latter study did not focus on patients hospitalized for respiratory failure. As our main question was to refine indications for short term lung transplant indication in CF children hospitalized in PICU, we focused on patients with respiratory failure due to pulmonary exacerbation and excluded rapidly reversible causes, as these conditions are associated with a better prognosis (8, 11).
In our study, main risk factors of mortality or lung transplantation during PICU were related to respiratory failure as shown by the level of pCO2 and Bicarbonate significantly more elevated in the group with pejorative evolution. Lower FEV1 and BMI, P.Aeruginosa and B.Cepacia colonization, mechanical ventilation during ICU did not reach significance probably because patients were already very severe and, most importantly, the study lacked power.
Death or lung transplant during the 36 months following PICU admission were mainly linked to chronic respiratory failure, as assessed by an increased risk in patients with home supplemental oxygen or non-invasive ventilation, and in those already on transplant list. This is also supported by the observation that patients with a poor evolution had increased levels of bicarbonate and decreased level of chloride at hospitalization which reflects chronic respiratory acidosis with increased kidney bicarbonate reabsorption and decreased chloride reabsorption (24). However a renal CFTR dependent participation in the regulation of bicarbonate serum level cannot be excluded (25). S.Maltophilia lung infection was significantly more prevalent in children with a bad evolution and may be an additional pejorative associated factor. S Maltophilia has been reported in association with low FEV1 in children, however, epidemiological data are still needed to confirm the pathogenicity of this microorganism (26). Finally, patients with home nutritional support were also at increased risk of death or lung transplant, probably reflecting the increased energy expenditure frequent in patients with CF and chronic respiratory failure (27). Each of these factors could be considered markers of more severe underlying diseases, highlighting the importance of longitudinal disease control and proactive early care. Similarly, Smith et al. also reported increased odds of mortality in patients with features of severe respiratory failure such as hemoptysis, pneumothorax, bacterial/fungal infections, malnutrition, and need for noninvasive or invasive respiratory support (17).
Upon arrival in PICU, higher levels of pCO2 and MIV were also associated with poor outcomes, as already described in adults, while children undergoing NIV were at significantly lower risk of pejorative outcome (19, 28). This suggests that NIV should be tried in a first line approach for those patients instead of mechanical ventilation. Malnutrition and any other comorbidities were not associated with increased mortality in our cohort. This is unexpected as prior studies did show a direct correlation between BMI and outcomes such as mortality and pulmonary function in CF patients (27). However, as a matter of fact, most of our patients had nutrition support and their BMI was within the normal range. Finally, female sex was also identified as a risk factor. This is in line with previous publications that reported significantly higher mortality in female subjects (29, 30), although it is not a consistent finding (31).
Children with CF enrolled in this study often received a lung transplant during their PICU hospitalization. Those admitted after July 2007 benefited from the implementation of the HELT system that prioritizes graft allocation to patients with short-term severe prognoses. This new organ allocation system was not associated with higher post-transplant mortality in adults, with a survival rate at 12 months around 80% (32, 33). Twelve children of our cohort were transplanted within the frame of HELT and their survival at 12 months was comparable to that of adults.
Most importantly, our results show very low rates of survival without lung transplantation (17%) during the 36 months following PICU admission for children with chronic respiratory insufficiency. Lung transplantation should therefore be considered early after PICU admission and children referred to lung transplant centers immediately after their discharge. A study conducted in France found that 80% of CF patients who died between 2007 and 2010 were not on any lung transplantation waiting list, and that this lack of listing was primarily related to late or lack of transplantation referral, rather than contraindication to transplantation (34). Our study unveils a similar observation as only half of the children in our cohort were on the waiting list for lung transplantation. The CF Foundation recommends lung transplant referral in children and adolescents to be anticipated no later than when FEV1pp is < 40% = or FEV1 is < 50% and rapidly declining or associated with markers of shortened survival including hypoxemia, hypercapnia, pulmonary hypertension, malnutrition, multiple exacerbations, massive hemoptysis and pneumothorax (35). We state that children on enteral nutritional support, chronic oxygen or NIV at home should be referred for lung transplants well, all the more they are female and colonized with S.Maltophilia.
This study has several limitations. Its retrospective design precludes conclusions regarding causality between the factors identified and the risk of mortality or lung transplantation. The small number of patients limits the statistical power of the analysis conducted. The external validity of this monocentric study needs to be confirmed in other centers, especially since our center is as a referral hospital for CF patients registered on the lung transplant waiting list and therefore admits the most severe patients from several regions in France. Finally, within such an extended timeframe, changes in CF care such as implementation of extracorporeal membrane oxygenation (36), and the advent of CF modulators (37), may jeopardize the data.