Data obtained in the studies evaluating automatic adjustment of FiO₂ have also illustrated the effort required to maintain oxygenation within the intended range. As expected, the number of FiO₂ adjustments by a dedicated caregiver was greater than those made by the routine caregivers and was also more effective in maintaining oxygenation within the intended range (Bhutani et al. 1992; Urschitz et al. 2004).

The effort spent in maintaining SpO₂ within an intended range also depends on the frequency of fluctuations. As expected, those studies involving infants with frequent episodes of hypoxemia documented more frequent manual FiO₂ adjustments. During routine care the caregiver must also be attentive to wean FiO₂ to the basal level after a hypoxemia spell has resolved to avoid rebound hyperoxemia. This is commonly not done consistently and soon enough during routine care. This is illustrated in Fig. 60.2 where a routine caregiver was more attentive to intervene during episodes of hypoxemia than to prevent hyperoxemia. The gradual reduction in basal FiO₂ during the automated period is accompanied by more frequent but milder episodes of hypoxemia.

Continuous and accurate availability of the infant’s oxygenation is key in achieving an effective and safe automated adjustment of FiO₂. The accuracy of pulse oximetry, the most commonly method used for automated FiO₂ adjustment, can be affected by motion artifact. Poor accuracy can result in unnecessary exposure to supplemental oxygen levels during periods of falsely low SpO₂ or prolonged hypoxemia if this goes undetected. These risks are present during routine care as well as during automated FiO₂ adjustment. Other factors such as poor perfusion or inadequate probe placement and function can also reduce accuracy (Bucher et al. 1994). In general, conditions that impair SpO₂ reliability and make its use questionable during routine care for an individual infant should also be considered as contraindications for its use for automated FiO₂ adjustment.

The reliability of pulse oximetry in correctly identifying hypoxemia is important in automated FiO₂ adjustment. Although most reports indicate reliable detection of hypoxemia by new pulse oximeters (Bohnhorst et al. 2000; Hay et al. 2002), some hypoxemia alarms during periods of infant activity are considered false due to the possibility of movement artifact. However, true hypoxemia spells can be associated with or result from the infant’s activity. These spells have been linked to increased activity and are more frequent when infants are awake (Bolivar et al. 1995; Esquer et al. 2007; Dimaguila et al. 1997; Lehtonen et al. 2002). This is important because artifactual hypoxemia spells during automated FiO₂ adjustment can result in unnecessary oxygen exposure. Reassuringly, in infants with frequent hypoxemia spells, automatic FiO₂ adjustment resulted only in minimal hyperoxemia overshoot (Claure et al. 2001, 2009). One would expect a high rate of hyperoxemia overshoot if most of the hypoxemia episodes were due to artifact.

Reduced attentiveness and poor observation of respiratory status may be an unwanted consequence of automated FiO₂ adjustment. In some instances additional oxygen may only provide palliative support while not addressing the root cause of hypoxemia such as hypoventilation. To avoid this, automated systems must warn the caregiver when the automatically set FiO₂ remains consistently elevated to maintain oxygenation in the desired range. Although this can be a potential drawback, it can be argued that the automatic increase in FiO₂ is a necessary first step to maintain the infant properly oxygenated and reduce exposure to prolonged hypoxemia if the clinical intervention is delayed. Another important issue is the acceptance and adaptability of this form of automated support to the busy clinical environment of a newborn intensive care unit. This has been recently assessed under routine clinical conditions over 24-h periods without any adverse events (Claure et al. 2011).

The optimal range of oxygenation for preterm infants has not been determined. Exposure to high oxygenation levels has been shown to worsen respiratory and visual outcome with minimal or no beneficial effects. Observational studies where lower target ranges were used to avoid the effects of extremely high SpO₂ suggested improved outcome (Tin et al. 2001; Chow et al. 2003; Wright et al. 2006; Vanderveen et al. 2006), but it is unknown how well those ranges were maintained and is possible the actual SpO₂ may have exceeded those ranges. More recently, important evidence from a large trial indicated increased mortality with a lower SpO₂ target range (SUPPORT Study Group of the Eunice Kennedy Shriver NICHD Neonatal Research Network et al. 2010). Because automatic FiO₂ controllers will target a range set by the clinician, it is recommended that selection of the target range be done with caution and considering the possible physiologic effects of a more effective maintenance of a desired target range of SpO₂ by the automatic FiO₂ controller.

In summary, automated adjustment of FiO₂ can improve maintenance of oxygenation within the intended range, reduce hyperoxemia and supplemental oxygen, and reduce staff workload. Larger clinical randomized trials are necessary to determine the effects of this automated form of support on survival and long-term respiratory, ophthalmic, and developmental outcome.