With advancements in medical technology, an increasing number of patients are surviving in the pediatric intensive care unit (PICU). However, some of these patients may not fully recover and may require life support equipment, such as mechanical ventilators, for extended periods. This presents a significant challenge for critical care providers and families worldwide(1–3). In developed Western countries, children who are dependent on technology often receive care from home care nurses and social workers, along with support from the community and government. In contrast, this model of child life support has not been widely implemented in mainland China.

In China, critically ill children who are still on ventilators but in relatively stable condition are typically transferred to national or regional children's medical centers for further treatment. However, none of these PICUs permit family members to accompany their children, and there are no alternative units available to care for these ventilator-dependent patients. The PICU is the only facility where families can remain with their children. Some families may opt to discontinue treatment due to financial constraints. However, an increasing number of families are choosing home ventilation to facilitate their children's reintegration into family life and society.

However, there is limited data available regarding patients requiring prolonged mechanical ventilation (PMV) nationwide, and follow-up studies are scarce. In this context, our multicenter collaborative group was established to further investigate medical management strategies and outcomes for children with PMV. This includes examining factors such as mechanical ventilation modes and settings, physiotherapy, tracheostomy, and other interventions, with the aim of improving outcomes for this population.

This is a retrospective study examining ventilation strategies for weaning patients with PMV across multiple national and regional children's medical centers in mainland China, utilizing a representative sample of cases. The definition of PMV in children varies(11); however, we have chosen to use the NAMDRC definition for comparison. Among the patients with PMV, 14.6% (46 out of 314) died in the PICU, which is comparable to the 15.9% mortality rate reported in China from 2017 to 2019 (12), but significantly lower than the 29% reported in a previous review(2). This lower mortality rate may be attributed to the fact that these ventilator-dependent patients received continuous intensive care in the PICU.

We included only survivors in our study of ventilation strategies for weaning. Excluded population (died/withdraw)'s characteristics was showed in Supplementary Table S1. Our findings indicated that 58. 1% (136 out of 234) of the patients were successfully weaned from mechanical ventilation prior to their discharge from the PICU. This result is comparable to the 50% weaning rate observed before discharge from the hospital in a meta-analysis of adult patients(13).

Among all study populations, the highest proportion of cases requiring PMV was attributed to central nervous system (CNS) diseases (41.5%), followed by lower airway diseases (31.2%). In European and American countries(14, 15), the proportion of delayed withdrawal of treatment due to neuromuscular diseases and congenital metabolic disorders was the highest. However, in China, due to unique social and medical conditions, a significant number of patients with congenital diseases opted for immediate withdrawal of treatment following diagnosis.

Compared to China's data from 2017 to 2019 (12), the proportion of PMV attributed to lower airway diseases exhibited a downward trend, decreasing from 41.6% to 31.2%. This decline may be attributed to several factors. First, the COVID-19 pandemic prompted the implementation of stringent national control policies, including mandatory mask-wearing, which contributed to a reduction in the overall incidence of respiratory infectious diseases. Second, advancements in the diagnosis and treatment of respiratory diseases, along with an improved understanding of mechanical ventilation and programmed weaning strategies, have facilitated earlier extubation for many children suffering from lower airway diseases. Additionally, many children's hospitals have established separate cardiac intensive care units (CICUs), which were not included in our study; consequently, the proportion of cardiovascular diseases accounted for only 4.7%.

We surveyed the ventilation strategies employed for patients with PMV and compared our findings to those from several years ago (12). We discovered that all patients received invasive mechanical ventilation, except for one patient who only received non-invasive ventilation. Most units preferred control modes, particularly pressure control, while only two of the eleven units opted for SIMV mode as the initial choice. By the 21st day of ventilation, 47% of patients remained on control mode. When considering weaning from mechanical ventilation, the strategies employed included SIMV, PSV, and CPAP. Prior to extubation, only 34.6% of patients underwent a SBT. This low percentage may be attributed to the high prevalence of central nervous system and neuromuscular diseases among the patients. Non-invasive ventilation was frequently utilized to prevent or manage post-extubation respiratory failure in high-risk patients. Due to the retrospective nature of this study and the lack of standardized weaning protocols across different units, we were unable to determine which approach was superior to the others. NAVA has been implemented in recent years, and research in adults has shown that NAVA improves diaphragm efficiency, unlike PSV (16). However, only two patients in our study received NAVA. The advantages and disadvantages of NAVA, as well as its potential to shorten ventilation time, still require further investigation in pediatric patients.

Early mobilization and physiotherapy(17) are crucial adjunctive therapies that play significant roles in successful weaning. The ATS/CHEST guidelines recommend initiating mobilization as early as 24 h after ventilation in adults (18). During ventilation, 63.2% of patients received physiotherapy, with the majority undergoing passive limb movements; only five patients participated in active exercises. Additionally, the case report form we designed did not capture the initial timing and frequency of mobilization, which hindered our ability to identify differences between the weaning success or not. However, in the tracheostomy group, the proportion of patients receiving physiotherapy was considerably higher, likely due to 87.3% of these patients having central nervous system and neuromuscular diseases. In the absence of physiotherapists, these patients were more likely to receive physiotherapy.

Since a strong cough can predict extubation success and may reduce ICU length of stay and mortality (19), cough augmentation techniques are employed to facilitate extubation (20). However, the results indicate that these techniques seem to have no effect on weaning.

In our study, tracheostomy was performed on 26.9% of the patients requiring PMV, which is comparable to Cinotti's report (21) of 25.6%, although they defined PMV as lasting more than 10 days. Half of the tracheostomies (32 out of 63) were conducted in two PICUs in Shanghai. In China, parental decision-making regarding long-term ventilation is influenced by various factors, including national healthcare policies, economic conditions, uncertain prognoses, and psychological stress. These factors complicate the decision-making process for parents (22). In Shanghai, an international metropolis, both healthcare providers and family members demonstrate a higher acceptance of tracheostomy compared to other regions in China. Additionally, Shanghai benefits from relatively higher economic levels, better medical insurance, and stronger community support, contributing to the increased proportion of tracheostomy procedures performed.

The average time to perform a tracheostomy was 29 days after the initiation of mechanical ventilation, significantly longer than the 10–12 days reported for adults (23, 24) and also longer than our data from 2017 to 2019, which indicated an average of 21 days (12). In pediatric patients, this duration is considered a late tracheostomy, regardless of the definitions of early tracheostomy, which are defined as occurring before 7 days of intubation according to Holscher et al. (25) or within 14 days as per Lee et al. (26) Both of these studies suggested that early tracheostomy reduces the use of sedatives and shortens the length of stay in the PICU, although they found no impact on the incidence of hospital-acquired pneumonia. In our study, all tracheostomies were classified as late, and the length of stay in the PICU was significantly longer for the tracheostomy group. However, among tracheostomy patients, those who underwent the procedure within three weeks had a shorter PICU stay compared to those who had late tracheostomies. This discrepancy may be partly attributed to epidemic prevention and control policies that limited the flow of patients, as well as the absence of specialized units for stable patients. Families of these patients spent considerable time in the PICU to receive training in necessary knowledge and skills before returning home or to rehabilitation centers.

Our study suggests that the causes of ventilator dependence is associated with weaning outcomes. Central nervous system (CNS) diseases were identified as the most common cause of PMV. In comparison to this subgroup, the chance of weaning success in patients with lower airway diseases had one more times. It is understandable that CNS diseases require a longer recovery time. On the day of PMV diagnosis, each 10% increase in FiO₂ was associated with a 33% increase in the risk of weaning failure.

As a retrospective study, our data collection from case records is inadequate. Nevertheless, this study can represent the current status of children on PMV in mainland China. There are still unresolved questions that need to be addressed through well-designed multicenter clinical randomized controlled trials (RCTs) examining various protocols for weaning strategies, physiotherapy, tracheostomy, and other related interventions. Specialized weaning centers should be established by governments and healthcare providers. Our study population was restricted to children who survived to undergo weaning assessment. By excluding non-survivors (a group with inherently higher illness severity), our identified predictors of weaning success are specifically applicable to this survivor cohort and should not be generalized to all children who develop prolonged mechanical ventilation. Our assessment of sedation and analgesia was limited to a binary classification (“any use”). We lacked granular data on medication type (e.g., analgesic vs. sedative), dosage, depth of sedation, or specific treatment targets. This heterogeneity limits precise interpretation of its association with weaning outcomes.

In our multicenter retrospective study, the characteristics of pediatric patients requiring PMV who survived in the PICU showed minimal variation. The most common cause of ventilator dependence was central nervous system diseases. The proportion of successful weaning in the ICU was low, and the timing of extubation and tracheostomy varied significantly. Among survivors, patients with lower airway diseases are easier to wean, while a higher FiO2 on the day of PMV diagnosis was associated with not weaned by the time of PICU discharge.