Unplanned extubation (UE), i.e., the unintended dislodgement of the endotracheal tube (ETT) (1) is a serious adverse event in neonatal intensive care with complications such as bronchospasm, bradycardia, and prolonged subsequent mechanical ventilation (2). The emergency reintubations that often follow has been associated with both a higher number of intubation attempts, and a higher rate of airway injury (3).

Compared to the pediatric and adult intensive care populations, UE is more frequently observed in neonates (4), likely due to a less frequent use of sedation (5), shorter trachea, and the use of uncuffed ETTs. Accordingly, reported risk factors for UE include infant agitation, poor ETT fixation, bedside procedures (e.g., suctioning or re-taping of the ETT) (2), and high ETT positions (6). A review by Silva et al. report a median incidence of 1.98 (IQR: 0.14–5.3) per 100 ventilator days (2).

Skin-to-skin care (SSC) of preterm infants is a mode of care with several important advantages for both the infant and parents. Compared to conventional care (CC) in a crib/incubator SSC has been shown to reduce mortality (7), lessen the risk of infection and hypothermia, improve physiological stability (8), and reduce pain response (9). While SSC is increasingly being applied also in a high-tech environment in the care of infants born extremely preterm (10), the impact of SSC on the incidence of UE risk has to our knowledge not been previously investigated.

We aimed to study the incidence of UEs in a setting where SSC is extensively applied also during mechanical ventilation, and compare the incidence of UEs during SSC and CC. We also aimed to evaluate the short-term respiratory outcomes after UE.

Of all infants admitted during the study period 2021–2023 (n = 1935), 603 infants received mechanical ventilation for a total of 3,353 days (Figure 1). Sedatives and/or analgesia are not used routinely in infants receiving mechanical ventilation in our unit, but all infants are assessed regularly for pain and stress with observational pain scores (Neonatal Pain, Agitation and Sedation Scale) and vital signs, if need for sedation/analgesia would arise. The standardized route for intubation was the oral route (n = 128), four patients were nasally intubated. The total number of UEs during the study period was 132 in 70 patients. Among these 70 patients 38 had one UE, and 32 had more than one UE (maximum 7 UEs in one patient). Eighty-two (62%) and 50 (38%) of UEs occurred during CC and SSC, respectively. The infants were predominately born extremely preterm (<28 weeks gestation) (Figure 2), their demographical characteristics are displayed in Table 1. No difference in ventilatory settings or FiO₂ were noted between CC and SSC prior to UE. The mean daily duration of SSC in all infants on mechanical ventilation during 2021–2023 and irrespective of UE, was 2.67 h (range 1.50–4.98 h).

The present study presents data on the incidence of UE in neonatal intensive care depending on whether the infant is receiving SSC or CC. In the studied cohort, UE was several-fold more frequent during SSC than CC. The overall incidence of UE per 100 ventilator days over the 3-year period was 3.9, comparable to previous studies. A review by Silva et al. reported a median incidence of 1.98 (IQR: 0.14–5.3) per 100 ventilator days (2), while others studies reported rates ranging from 2.1 (11) to 3.3 (12), and 4.75 (13). A benchmark target of less than 1 UE per 100 ventilator days has been suggested by Silva et al. (2), as a realistic aim of quality improvement. The infants in our cohort had lower gestational age at birth than in the studies mentioned previously (GA 24.3 vs. 27.3, 27, and 26.7) (11–13), which might have led to a longer time on mechanical ventilation and our relatively higher UE incidence.

To our knowledge, this is the first study reporting the UE incidence during different modes of care, together with the time spent in the respective modes. The two previous studies that have reported a lower proportion of UEs occurring during SSC (12, 14) than in our study, did not take into account the length of mode of care, making absolute comparisons with our study difficult. Clearly, an approach where infants are only briefly cared for in SSC will demonstrate a low proportion of UE occurring during this mode of care. For feasible future comparisons we would like to propose applying the number of UEs per 100 ventilator days as the preferred measurement.

Seventy-two percent of the patients required reintubation within 120 min after the event, and while previous studies have reported lower reintubation rates (2, 3, 15), those cohorts included infants with several weeks higher median gestational age which may have influenced the outcome. Given that reintubation was not required in 28% of events, this implies that these infants could have been extubated earlier, under controlled circumstances, suggesting an opportunity to improve the identification of extubation readiness. The reintubation rates and ventilatory settings were similar when comparing SSC and CC. Oxygen requirements were found to be slightly higher after UEs that required reintubation which further strengthens the need to avoid such a stressful and potentially harmful adverse event (16).

The use of uncuffed tubes and the low use of sedatives during prolonged MV may influence the high incidence of UEs in our NICU (5, 17). In the present study almost all ETTs were uncuffed and placed using the oro-tracheal route according to standardized routine. While it has been argued that the nasal route allows a more secure ETT fixation (17), it has not been demonstrated to affect the rates of UE (18). Previously reported risk factors for UE, including poor ETT fixation, bedside procedures (e.g., suctioning or retaping of the ETT) (2), and high ETT positions (6), are also potential areas to analyze to identify patterns of UEs, and thereby areas of improvement.

During the three-year study period, we observed a decreased incidence of UE during SSC. While we cannot demonstrate any cause-effect relationship given the observational nature of our investigation, a quality improvement (QI) initiative was launched in November 2021, consisting of monthly feed-back reporting of the UE incidence on a staff lunchroom information screen as well as raising awareness about these highly unwanted adverse events at repeated staff meetings. As previously documented (5) it is conceivable that this might have contributed to the observed decline in UE, particularly during SSC. Nevertheless, the high rate of UE during SSC is of concern and should be further targeted in future QI efforts.

We believe that the observed increased risk of UE during SSC should not be taken as an argument to withhold or postpone SSC, given its many advantages such as physiological stability (19). The postnatal age at UE did not differ in SSC and CC (14 vs. 15 days), implying similar clinical settings at UE. SSC was initiated after 4–6 days of life for 22–24 GA, after 1 day for 25 GA, after first hours of life for 25–26 GA, and immediately after birth for >26 GA in the present cohort. Rather than refraining from SSC, optimization of routines for ETT fixation and positioning of the ventilator tubing, transfer procedures, and raising staff and parental awareness should be further pursued in order to enable both a lowest possible rate of UE, and parent-infant closeness.

In summary, the UE incidence was generally high and affected oxygen requirements after reintubation, further strengthening the adversity of the event. When correcting for time spent in each mode of care, expressed as UEs per 100 days of mechanical ventilation, the UE incidence was higher during skin-to-skin care than during conventional care. Future QI initiatives are necessary to enable both a low risk of unplanned extubation, and parent-infant closeness.