Table 3 summarizes the healthcare resources used among the children with prolonged admission to the PICU. A full 97.6% of the patients required invasive mechanical ventilation, 12% required high-frequency ventilation, and 65% required noninvasive ventilation. In turn, 33 patients (39.8%) required extrarenal filtration techniques: in all these cases venovenous hemodiafiltration was performed, with added peritoneal dialysis in 8 patients (9.6%).

A total of 86.8% of the patients required inotropic drugs in continuous perfusion (adrenalin in 75.9%), 22.9% were subjected to ECMO, and 4.8% required mechanical ventricular assistance. A total of 24.1% of the patients underwent heart transplantation (15 of them after admission to the PICU).

Bladder catheterization was carried out in 94.9% of the cases, with central venous accesses and arterial catheters in 92.8% and 92.6%, respectively. In turn, 90.2% of the patients required blood product transfusions during admission.

Three patients (3.6%) presented a tracheostomy and four a (4.8%) gastrostomy at the time of admission to the PICU. During their stay, another 13 patients (15.7%) underwent tracheostomy and 9 (10.8%) gastrostomy.

Infection was suspected in 32.1% of the patients at the time of admission to the PICU; the percentage corresponding to infections of nosocomial origin could not be established. A total of 96.3% of the patients suffered at least one nosocomial infection during admission—the most frequent condition being pneumonia (72%), followed by urinary infections (34.6%) and catheter-related infections (28.4%).

A total of 19 of the 83 patients with prolonged admission died. Mortality among the children with prolonged admission was significantly greater (22.9%) than in the rest of the patients (2%) (p<0.001). The most frequent causes of death were withdrawal or limitation of therapeutic effort (42.1%) and the non-adoption of resuscitation measures (10.5%). A total of 8.4% of the patients with prolonged admission had suffered cardiorespiratory arrest (CRA) before admission to the PICU, and 35.4% suffered CRA during their stay.

At the time of death, all the patients were subjected to mechanical ventilation (conventional mechanical ventilation in 89.5%, noninvasive ventilation in 5.3%, and high-frequency ventilation in 5.3%,); 89.5% were receiving vasoactive drugs in continuous perfusion; and 26.7% were subjected to ECMO.

There was no difference in mortality between patients under one year of age (23.3%) and the older children (22.6%). Table 4 depicts the relation of other variables to mortality. Although congenital heart disease, admission after surgery, ECMO, continuous extrarenal filtration techniques (CEFT), and blood product transfusions were more frequent in the patients who died, the differences only reached statistical significance in the case of CEFT. The patients who died presented a comparatively longer duration of mechanical ventilation, inotropic drug treatment and ECMO (Table 4).

Discharge of the survivors was to the pediatric ward (89%) or to other intensive care units (4.6%), while one patient (1.5%) was directly discharged home. Three patients (4.6%) were still in the PICU at the end of the study.

We have observed important healthcare resource utilization among the patients with prolonged admission, in coincidence with the findings of earlier studies.1,5,11,12

Tracheostomy and gastrostomy can improve management, reduce the complications and improve the quality of life of patients with prolonged admission.13–16 Although early tracheotomy is advocated in adults,17,18 in children the best timing for this technique has not been established. The mean intubation time prior to elective tracheostomy in children subjected to prolonged mechanical ventilation varies from 31 to 51 days, depending on the data source.14,15 In a survey of 63 pediatric intensivists in Canada regarding the use of programmed tracheostomy in children subjected to mechanical ventilation, 51% of the interviewed physicians considered programmed tracheostomy to be underused in children.14 In our study, the procedure was only carried out in 15.8% of the patients.

We recorded a high incidence of nosocomial infections,19–21 particularly respiratory infections probably related to the important use of invasive mechanical ventilation, in coincidence with the findings of earlier studies.22.23 Likewise, the prolonged use of central vascular accesses and bladder catheterization favors infection in these locations. Specific measures are therefore needed to reduce the duration of invasive techniques, and to prevent and quickly diagnose and treat nosocomial infectious processes in these patients.5,6,10,11,24–27

Most of our patients were discharged to other hospitalization areas. Patients with prolonged admission to the PICU often remain dependent upon technological systems—fundamentally mechanical ventilation—and so require specific care measures not available in the hospitalization wards.28 This fact often delays discharge from the PICU, despite the achievement of sufficient patient stability.

The creation of intermediate care units or units for technology-dependent patients, with sufficient personnel and specific material resources to safely treat subjects in clinical conditions that are unlikely to change (apart from the development of complications), could allow for more efficient healthcare resource utilization.29,30

Home mechanical ventilation facilitates the out-hospital care of some patients, increasing their quality of life and that of their relatives. However, this option requires the existence of care teams to conduct regular controls and to train the caregivers, and the family in turn must receive sufficient sociosanitary support. The scarcity of these aids often represents an obstacle for discharging patients from the PICU, contributing to further their chronification and institutionalization.28,31,32

Mortality among patients with prolonged admission is much higher than in the rest of the patients admitted to the PICU. In effect, the mortality rate in our series was 10 times greater—in coincidence with the findings of other authors in pediatric patients.2,3,33 It is important to mention that our PICU does not often admit patients for short-duration techniques or procedures—the low associated mortality of which would have contributed to lower our overall mortality figures.

The early identification of patients with a high mortality risk could allow more efficient use of the available resources, attempting to avoid the use of futile invasive procedures.7 Some factors, such as heart diseases, or the need for extrarenal filtration techniques or ECMO, are associated with increased mortality, though in our study most of them showed no statistically significant differences. Nevertheless, it must be taken into account that most of these factors are only indicators or markers of patient severity.

In our study, and in the same way as in other series,3,34,35 the cause of death in almost one-half of the cases was the withdrawal or limitation of therapeutic effort. Despite this fact, in most cases mechanical ventilation or vasoactive drug treatment was maintained up until death. This may be due to the way in which the limitation of therapeutic effort is carried out in our PICU, though it could also imply the excessive prolongation of techniques that may be regarded as futile. It therefore seems necessary to optimize the approach to such patients when death draws near—addressing and planning the measures to be taken, with a view to avoiding inadequate treatment prolongation.3–5,8,34

Considering the retrospective design of our study, it is important to point out that its main limitation is the absence of comparisons between the group of patients with prolonged admission and those without prolonged admission—a situation that precluded a multivariate statistical analysis of the risk factors for prolonged admission.

On the other hand, our PICU has a series of specific characteristics due to the high percentage of patients admitted with congenital heart disease. Other studies involving PICUs with more general patients are therefore needed to confirm the results obtained.

In conclusion, although children subjected to prolonged admission represent a small percentage of the patients seen in the PICU, they suffer important morbidity–mortality, account for much of the occupation of the unit, and involve important healthcare resource utilization. Early identification is required of those patients with a high risk of requiring prolonged admission, in order to establish specific management strategies designed to allow more efficient care and use of the available resources. The creation of intermediate care units for the management of these patients may help reduce resource consumption and increase the quality of the provided healthcare.