Monitoring Oxygenation and Gas Exchange in Neonatal Intensive Care Units: Current Practice in the Netherlands

Background Although recommendations in oxygenation and gas exchange monitoring in the neonatal intensive care unit (NICU) are available, little is known of the current practice. Aim To evaluate the current practice in oxygenation and gas exchange monitoring of the NICUs in the Netherlands. Methods An online survey-based questionnaire concerning preferences and current practice of monitoring oxygenation and gas exchange was sent out to all 107 neonatal staff members (neonatologists, neonatal fellows, and physician assistants) of the 10 NICUs in the Netherlands. Results The response rate was 42%. Pulse oximetry (PO), partial pressure of oxygen in arterial blood gas (paO2), and oxygen saturation in arterial blood gas (saO2) was used by, respectively, 100, 80, and 27% of the staff members for monitoring oxygenation. Of all staff members, 76% considered PO as the best parameter for monitoring oxygenation, 22% paO2, and 2% saO2. Blood gas, transcutaneous gas monitoring, endotracheal gas monitoring, and near-infrared spectroscopy was used by, respectively, 100, 82, 40, and 18% of the staff members for monitoring gas exchange. During endotracheal ventilation, 67% of the caregivers would exclusively accept arterial blood gas for gas exchange monitoring. In contrast, during non-invasive ventilation, 68% of the caregivers did not prefer arterial or capillary blood gas (CBG). CBG is found reliable in infants with warm extremities by 76% of the caregivers. Venous blood gas would be accepted by 60% of the caregivers, independent of the mode of respiratory support, and only when venous blood sample was needed for other reasons. Conclusion This survey identified a wide variation in preference in monitoring oxygenation and gas exchange monitoring among Dutch neonatal staff members.

Although all studies reported the advantages and disadvantages of the different methods and methods are recommended, the choice of the caregiver on what to use in daily practice is also influenced on the availability, practicality, and local preferences. Guidelines are needed to aim for a uniform policy, but first it is important to know how the current daily practice corresponds with the available recommendations. To investigate this, we performed a survey to inventory the current practice in monitoring oxygenation and gas exchange in all NICUs in the Netherlands.

MaTerials anD MeThODs
We sent a questionnaire to the 107 neonatal staff members [84 neonatologists, 18 neonatal fellows, and 5 physician assistants/ nurse practitioners (PA/NP)] working in the 10 Dutch tertiary NICU centers in 2012, who are registered by the Dutch Association of Pediatrics. The questionnaire included questions concerning the daily practice in monitoring oxygenation and gas exchange (see Table 1). The units in the Netherlands are all up-to-date level III NICUs, and there are very little difference in resources and employees. For this reason, we aimed for a national overview. Most equipment for monitoring is available in all Dutch NICUs. We aimed for at least three responders per NICU, a first general reminder was sent after 4 weeks, and the last reminder was sent specific to NICUs when the response rate was less than three. The local institutional ethical review board approved the questionnaire study.

statistical analysis
All replies were collected in MS Excel 2010 and expressed as percentages of total participants.

resUlTs
From the 10 NICU centers, 45/107 (42%) neonatal staff members responded and returned the questionnaire; of which 35 (42%) neonatologists, 8(44%) neonatal fellows, and 2(40%) PA/NP. From each unit at least three staff members responded. All staff members indicated that there were no specific guidelines available in the unit concerning monitoring oxygenation and gas exchange.

Blood gas
The type of blood gas (ABG, CBG, or VBG) used depended on the ventilation modus of the infant (Figure 1). In all units, most often ABG is used in intubated and mechanically ventilated infants, while in non-invasive ventilated infants, both ABG and CBG were preferential. Staff members considered the reliability of CBG depended on the presence of warm extremities (76%), capillary refill time <2 s (44%), pink extremities (37%), and normal central body temperature (22%). Overall, staff members considered CBG more reliable than VBG (Table 5); nevertheless, 11% uses VBG in ventilated infants and 13% in non-invasive ventilated infants (Figure 1). VBG and CBG were not considered reliable at all by 13 and 4% of the caregivers, respectively ( Table 5). In contrast, this opinion changes when a blood gas was due when venous blood sampling was simultaneously needed; the majority (60%) of the clinicians would then accept a VBG.

DiscUssiOn
This study provided an overview of the current clinical practice. Although most recommendations concerning monitoring gas exchange are followed, there is a variance in and between units in opinion about reliability and interpretation of the measurements. Oxygenation and gas exchange were most often monitored by PO, tcCO2, and ABG, while other current available measurements are infrequently or rarely used. Factors like practicality, personal experiences and aiming for less cumbersome measurements probably influenced current practice in monitoring. The response rate was low (42%), but this is higher than the average response rate of online voluntary survey in physician specialists (33). We anticipated a low response, and we therefore aimed for at least three responders of each NICU. We reached this target with a mix of inexperienced and experienced users. We believe that the respondents give a good reflection of the Dutch neonatal staff members.
According to the literature, the best method for monitoring hypoxia and hyperoxia is an ABG (analyzed with a co-oximetry unit), but we observed that staff members infrequently used the oxygen saturation in the ABG (27%). In addition, most caregivers were not aware whether the blood gas analyzer uses co-oximetry or is an estimation. While co-oximetry would be more accurate than PO; caregivers considered PO as the best and most commonly used monitoring for oxygenation. PO is non-invasive, non-cumbersome, easy to apply, and will provide continuous measurements. However, oxygen saturation can be overestimated by PO (7) and is inaccurate at oxygen saturation levels <70% and in hyperoxia >95% (29,34). The other disadvantages of the PO are motion artifacts, even in the motion resistant types, and low perfusion quality disturbances depending on probe site placement (29). While studies recommended continuous tcO2 monitoring for preventing hypoxemia/hyperoxemia (15,32), few caregivers (7%) used this. In contrast, most caregivers considered tcCO2 the best non-invasive continuous monitoring of CO2, which is consistent with a recent German survey (35). Studies demonstrated that tcO2/CO2 is well correlated with paO2/CO2, but tcO2 often underestimated paO2 while tcCO2 overestimated paCO2 (12, 15-17, 32, 35-37). In contrast to the literature, most caregivers considered tc measurements not reliable and is possibly mostly used for trend monitoring. It is possible that the frequency of replacement and calibrations has influenced the caregiver's experience in the reliability. While in the performed studies, the sensor is frequently replaced and calibrated, in daily practice this is not always possible. In addition, dislocated adhesive rings, skin burns, and decubitus can discourage caregiver in using tc monitoring. In this view, the preference of staff members to use tc monitoring in preterm infants was surprising, especially considering the fact that correlation is poor in preterm infants when compared to term infants (32,38). End-tidal CO2 monitoring is not frequently used in the Netherlands, most probably because the inaccuracy due to high respiratory rate, the low-tidal volume, and the increase of dead space volume (19).  The staff members considered ABG analysis the gold standard for monitoring gas exchange, which is consistent with the literature (23,31,36). The preference of performing ABG in ventilated infants probably reflects that most of the ventilated infants have an arterial catheter for hemodynamic monitoring, while CBG is well accepted in non-invasive ventilated infants. The reliability of CBG can be poor by disturbed perfusion, which is more likely to occur in the more critical and mechanical ventilated infants. The staff members considered the CBG for pH, pCO2, and sodium bicarbonate in a child with warm extremities reliable, which is consistent with the current literature (1,24,26,36).
The use of VBG is currently not recommended, as the studies performed are small and showed conflicting results (1,3,6,23,27,28). Nevertheless, most staff members will accept a VBG when venous blood withdrawing will occur, most likely due to practical considerations. However, so far there is no data whether VBG and CBG can be interchangeably used. Gillivray et al. concluded that VBG would be useful for decision making in the need for intubation, but not for monitoring ventilation (6).
cOnclUsiOn This survey identified a large variation in oxygenation and gas exchange monitoring in the Dutch NICUs. Although guidelines are absent in the units, recommendations are followed combined with practical considerations. In current daily practice, PO is considered the best method of monitoring oxygenation and blood gasses for monitoring gas exchange. The preference for the type of blood gas analysis depends on circumstances: type of respiratory support and the opportunity of simultaneously venous blood withdrawing. Tc gas monitoring is then considered the best noninvasive measurement for gas exchange but is infrequently used. These results can be helpful when developing a guideline that is evidence based as well as practical.

acKnOWleDgMenTs
We thank all participants for answering our questionnaire.