This observational study of 197 children on HMV represents one of the largest pediatric cohorts reported to date, and the first to describe experiences and trends in applying HMV over a period of 3 decades. The most striking change over time is the steep rise in the number of children in whom HMV was being applied, particularly during the last decade and in the youngest age group. Especially these children had very long stays in the PICU, both before and after HMV was initiated.
There are several explanations for the increase in the application of HMV in children. Presumably, the success of HMV in adults has led to a greater awareness among physicians involved in the care for children. Improvements in medical health care, increased patient autonomy and changing moral-ethical attitudes with respect to treatment decisions may explain the growing number of children admitted to and kept alive in the PICU . Improvements in technology have made HMV not only an alternative for death, but also succeeded in improving the quality of life for most of these children; consequently children are nowadays being referred for HMV more frequently .
We found an escalating need for PICU beds, especially for the youngest children (Table 3), and specifically these children consumed the most PICU admission days. The retrospective nature of our study limited an in-depth investigation of the reasons for these prolonged stays. The pre-HMV length of stay probably reflects the complexity of the medical condition in young children: most of them were admitted after an acute deterioration of their respiratory condition, often before the diagnosis or prognosis of the underlying disorder was made. This is underscored by the high incidence of non-elective initiation of HMV in this group. We have the impression that the most time post-initiation of HMV was required for fine-tuning ventilator settings, training parents and organizing professional home care. In the youngest children, with often less stable respiratory conditions, these steps are often more difficult and time-consuming than in older children.
The long post-HMV period in the youngest age group particularly raises our concern. Though lower per patient than in the UK  and the US , it amounted to a total of 1,351 post-HMV PICU days in the last decade, which means blockade of one PICU bed for about 4 months for each year in the last decade. During the last decade, a discharge coordinator was introduced in our Center for HMV, and efforts have been made to streamline the discharge process in order to shorten PICU admission. As shown in Table 3, this resulted in a decrease of post-HMV days per patient compared to the previous decade, but despite this, the total PICU length of stay rose considerably because of an increase in young patients starting with HMV. Since it is likely that this trend is going to continue, the need for PICU beds will probably continue to rise as well. It could be worthwhile to establish designated units outside the PICU for clinically stable children requiring HMV . Such a setting, which is less stressful and more focused on rehabilitation than on acute care, might not only be a better but also a cheaper and safe alternative [15–18].
The amount of literature on trends in the application of HMV in children is limited. We are aware of only one other study covering the same time span, but this study contained both children with and without HMV and had different end points . Other studies covered a shorter time span [10, 19, 20] or a single time point [7, 9, 21, 22], contained children with either invasive [13, 20] or noninvasive HMV , or were mainly focused on survival  or hospital admission days . In comparison to other studies, invasive HMV was used with a comparable frequency (38 %) as in Massachusetts (49 %) or Italy (41 %), but markedly more frequently than in Australia (22 %), the UK (23 %) or Turkey (32 %) [7–9, 21, 24]. Invasive HMV is our first choice for the youngest children as they have already been tracheotomized during their stay in the PICU and need to keep a tracheostomy for the time being to evacuate airway secretions. Moreover, noninvasive ventilation may cause mid-facial retrusion if applied in early childhood and is often poorly tolerated, particularly when applied for 24 h per day . Improvements in interfaces for noninvasive HMV and new developments such as mechanical insufflation/exsufflation have, however, caused a decline in tracheostomal ventilation, also for very young patients [26, 27].
At the end of our study, 11 children (6 %) were eventually weaned off of HMV. In other studies weaning varied from 9 to 39 % [8, 10, 13, 14, 24]. Mortality was considerable with 22 %, but nearly always related to the underlying disorder, and comparable to the mortality of 7–32 % found in other cohorts [8, 10, 13, 14, 20, 24]. The observation that most patients who died had used HMV for several years strongly suggests that HMV prolonged their lives.
Our retrospective study has several limitations. Being a single center study, our results may not be automatically applicable to other centers. However, the four Centers for HMV in the Netherlands have always been comparable with respect to their procedures, and we feel that our results reflect the Dutch situation. As HMV in the Netherlands is well organized, it is unclear to what extent our findings can be extrapolated to patients in other countries. Given the number of patients and time span involved we consider our results relevant for providing trends with regard to HMV in children and, thereby, for comparison with other studies.
In conclusion, in the past 3 decades the number of children, and particularly those aged 0–5 years, who received HMV in our center increased substantially. This was associated with an escalating need for pediatric intensive care beds. If this trend continues, the establishment of specialized facilities outside the PICU for clinically stable children requiring HMV should be seriously considered.