In the August issue of the Journal, a Clinical Opinion article entitled “Oxygen for intrauterine resuscitation: of unproved benefit and potentially harmful” by Hamel et al was published. The Journal encourages such opinion articles, and this one in particular raises potentially important questions about the common practice of administering oxygen to women in labor when the electronically monitored fetal heart rate (FHR) patterns suggest the possibility of fetal hypoxia.

There is no question that the practice of administering oxygen to women in labor is vastly overdone. Even before the Eunice Kennedy Shriver National Institute of Child Health and Human Development categorization of fetal heart rate patterns and now reinforced by the fact that more than 80% of women in labor at some point are in category II despite the fact that many of their patterns are not associated with fetal hypoxia (eg, variable decelerations with moderate variability), a significant number, if not the vast majority, of such women will receive oxygen at some point during their labor.

Administration of oxygen in response to a concerning FHR pattern certainly can and probably does create anxiety for the patient and also creates an environment in labor and delivery in which care givers can be incited to be more anxious about a given FHR pattern than is necessary, possibly adding to other unnecessary and unwarranted interventions.

We are strongly in favor of being more selective about which patients actually do have patterns consistent with hypoxia, as we described in a recent paper by Clark et al. Such patterns that will more likely benefit from oxygen administration include category III tracings while preparations are being made for delivery and category II tracings with recurrent late or prolonged decelerations, significant variable decelerations, bradycardia, or tracings with any decelerations with minimal or absent variability lasting 30-60 minutes.

Hamel et al appropriately point out that there are potential adverse effects of oxygen administration. We accept that in fetuses that are not hypoxic or not likely to be hypoxic based on FHR patterns, this common practice could theoretically do more harm than good. The controversy over this article, however, lies in the authors’ final conclusion that “until it is studied properly in randomized clinical trials … maternal oxygen supplementation in labor should be reserved for maternal hypoxia and should not be considered an intervention for nonreassuring fetal status.” The authors do acknowledge that studies are reasonably clear that maternal oxygen administration does improve fetal oxygen levels and ameliorate FHR patterns indicative of hypoxia such as late decelerations.

The authors’ arguments that oxygen may be harmful are based on several points. They cite literature that suggests that maternal oxygen administration may actually lower the fetal pH and may result in the creation of dangerous free radicals, in both mother and fetus, from hyperoxia or reoxygenation following periods of hypoxia. The data that increasing maternal pO ₂ results in a lower pH in the fetus are weak at best and in many ways are counterintuitive. Nonrandomized trials do not support this association.

In the 2 randomized trials of maternal oxygen administration during labor with reassuring FHR patterns to which the authors refer, there were no differences in the mean umbilical arterial pH at birth, although one study did report more babies with an umbilical artery pH of less than 7.2 in the group receiving maternal oxygen. One of the animal studies referenced involved a 30 minute administration of oxygen to laboring nonhuman primates with acidotic fetuses. There was only a minimal resultant improvement in pH identified, but with severe metabolic acidosis because of tissue hypoxia, 30 minutes is probably not an adequate duration of time for oxygen replacement to result in a reversal of metabolic acidosis.

The time for oxygen supplementation necessary for reoxygenation of cells, conversion back to aerobic metabolism, and clearing the saturated buffers of excess hydrogen ions often takes significantly longer than 30 minutes. Hamel et al did note that such therapy did improve fetal oxygen levels and decreased the frequency of late decelerations in the majority of these animal fetuses.

In terms of the issue of possible damage to the fetus from hyperoxia, the authors use the adverse effects of neonatal hyperoxia as an example of potential harm. Neonatal effects, and for that matter for essentially all viable tissue, there is no doubt that too much oxygen for too long can do harm. But in the neonate, levels of pO ₂ and/or SaO ₂ that result in harm far exceed those the fetus reaches with maternal oxygen administration.

In 2 studies using fetal pulse oximetry in laboring women with (which the authors did not reference in their Clinical Opinion article), administration of oxygen to mothers with normally oxygenated fetuses resulted in an average increase in fetal oxygen saturation of 5% when using a simple mask and 7-15% when using a nonrebreather face mask. However, in hypoxic fetuses (those with baseline oxygen saturations of 30-40% in which 40-60% is considered normal), there was a 20% increase in fetal oxygen saturation with a simple face mask and a 26-37% increase with a nonrebreather mask. None of these increases would be considered significant hyperoxia but did instead demonstrate return of an abnormal level of oxygenation to a normal level. Thus, these levels of increased oxygen saturation in both normoxic and hypoxemic fetuses are far less than seen with oxygen administration in the newborn.

Physiologically this makes sense because the maximum fetal pO ₂ cannot exceed that seen in maternal venous blood. The human placenta has been shown to behave as a concurrent blood flow pattern and thus the pO ₂ in the fetal umbilical vein reflects diffusion from the maternal venous side of the spiral artery-capillary-vein loop and not the arterial side. Thus, the maximum pO ₂ in the fetal umbilical vein cannot exceed maternal venous pO ₂ , which is considerably less than maternal arterial pO ₂ under either room air conditions or with maternal hyperoxia.

In terms of the adverse effects of free radical formation resulting from maternal hyperoxia, which Hamel et al raise as another concern, one wonders whether this is a realistic concern. So many individuals receive high-flow oxygen in ambulances, with anesthetics in surgery, and in other medical situations and it is not voiced as a concern by such specialists or demonstrated in studies of which we are aware to cause clinically apparent harm.

Finally, although we strongly disagree with the conclusions drawn by Hamel et al to limit oxygen administration only to hypoxic mothers and not for fetal concerns, we do believe it is important for the reasons described in this editorial to be more selective about which patients should have supplemental oxygen in labor and when administering it to do so in the most effective manner. We hope this report provides clinicians pause when giving oxygen to mothers whose FHR patterns are unlikely to be indicative of hypoxia. Beyond the fact that this results in no improvement of fetal status, the intrusiveness of the oxygen mask and the anxiety created in the mother when it is apparent that caregivers are concerned about her fetus could possibly or at least theoretically have adverse effects. Nonetheless, based on the knowledge that fetal hypoxia can be improved by maternal hyperoxia, to argue that we should not administer oxygen in labor complicated by a concerning FHR indicative of fetal hypoxia until randomized clinical trials demonstrate benefit is, in our view, a conclusion that goes too far and cannot be supported.