Instrumentation involving technology similar to adult pulse oximetry has been developed that might allow measurement of fetal arterial oxyhemoglobin saturation once the membranes have ruptured. Sensors to be placed transvaginally between the uterine wall and fetal face during labor have been designed that measure fetal arterial oxygen saturation. Although monitoring of fetal arterial oxyhemoglobin saturation has been associated with many technical difficulties,1 several investigators have reported the possibility of obtaining meaningful information on intrapartum fetal oxygenation. In 1992, McNamara and colleagues2 reported a significant correlation between fetal arterial oxyhemoglobin saturation values obtained just before delivery and umbilical cord blood gas results in 28 uncomplicated pregnancies. Using continuous pulse oximetry, Luttkus and colleagues3 described a significant association between fetal arterial oxyhemoglobin saturation values and results of simultaneous fetal scalp blood analyses, but others4 have reported poor correlation in pregnancies complicated by nonreassuring fetal heart rate (FHR) patterns, fetal growth restriction, or thick meconium.
Expanding on animal studies5 designed to establish arterial oxygen saturation thresholds at which metabolic acidosis develops, Dildy and co-workers6 suggested that a fetal arterial oxyhemoglobin saturation cutoff value of 30% would be reasonable for detection of clinical fetal hypoxemia. Recently, Goffinet and colleagues7 found that the 30% fetal arterial oxyhemoglobin saturation threshold was useful in detecting fetal compromise in 174 women with abnormal FHR tracings during labor, which suggests that it might be useful to measure fetal oxygen saturation during labor and that values less than 30% might indicate fetal compromise.
We recently participated in experimental testing of modified fetal pulse oximetry sensors (FS-14; Nellcor Puritan Bennett Inc., Pleasanton, CA) designed to improve signal registration times during labor (ie, the amounts of time from initial sensor placement that a saturation value is electronically measurable). Data collected during sensor testing were analyzed retrospectively, with particular emphasis on indices of fetal compromise, in relation to fetal arterial oxyhemoglobin saturation values above or below 30%. Short periods (seconds to minutes) of saturation below 30% were so common, occurring in 53% of the pregnancies studied, that they potentially precluded meaningful interpretation of fetal oxygen desaturation. Therefore, we tested the hypothesis that duration of desaturation, rather than simply its occurrence, might be more meaningful in predicting fetal compromise.
Materials and Methods
In accordance with a protocol approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and Parkland Memorial Hospital, an experimental, fetal arterial oxygen saturation sensor (Nellcor Puritan Bennett Inc.) was introduced into the uteri of consenting women by one research nurse (RGS) after spontaneous or artificial rupture of fetal membranes. Eligible women had uncomplicated singleton pregnancies at or beyond 36 weeks' gestation. Continuous electronic fetal monitoring was used, and selection of women was not prompted by preexisting concerns about FHR.
Equipment used in this investigation consisted of fetal oxygen sensors (IDE-FS-14; Nellcor Puritan Bennett Inc.; and modified prototypes) and a fetal arterial oxyhemoglobin saturation monitor (N-400; Nellcor Puritan Bennett Inc.), a conventional electronic fetal monitor (M1350B; Hewlett-Packard, Böblingen, Germany), and a computer interface capable of collecting data from the saturation equipment and fetal monitor. Fetal heart rate and uterine-contraction pressure data were sampled and recorded at 250-millisecond intervals. Fetal arterial oxyhemoglobin saturation data were sampled and recorded at 17.5-millisecond intervals. To facilitate analysis, those data were averaged in sequential (nonoverlapping) epochs of 10 seconds' duration.
Operation of the monitoring system was based on principles of spectrophotometry and plethysmography. The system included an optoelectronic sensor and a microprocessor-based monitor. The sensor had two low-voltage, light-emitting diodes as light sources and one photodetector. One light-emitting diode emitted red light (735 nm) and the other emitted infrared light (890 nm). When light from each light-emitting diode passed through fetal tissue at the sensor application sight, a fraction was absorbed. The photodetector measured the light that was not absorbed (ie, that was reflected), a process similar to measurement of transmitted light in conventional pulse oximetry. Because oxyhemoglobin and deoxyhemoglobin have different light-absorption characteristics, less red light is absorbed by oxyhemoglobin than by deoxyhemoglobin. Relatively more infrared light is absorbed by oxyhemoglobin than by deoxyhemoglobin. Pulse oximetry uses the ratio of those differences to determine fetal oxygen saturation during each arterial pulse.
Before placement of the saturation sensor, a vaginal examination was done to determine cervical dilation (3 cm or more), station (−2 or lower), and fetal position, to ascertain the optimal place for sensor insertion. The optimal site for sensor placement was on the cheek or temple of the fetal head. With the use of aseptic technique, the sensor was guided through the cervix and placed satisfactorily (Figure 1) using indicators that displayed sensor contact and signal quality on the N-400 monitor. The sensor was placed on the soft tissue of the fetal cheek. The research nurse, continuously attending each participant, did periodic sensor adjustments for maximal data acquisition. The sensors stayed in place until they were expelled spontaneously or delivery occurred. Fetal arterial oxyhemoglobin saturation information was not available to clinicians who were managing women enrolled in this study. Fetal scalp blood analyses were not done.
Statistical analyses of categorical data included Pearson χ2 for contingency tables and Fisher exact test. The Mantel-Haenszel χ2 was used to test for trends. Student t test was used for comparing means. P values (two-sided) less than .05 were considered significant. Analysis was done with SAS 6.12 statistical software (SAS Institute, Cary, NC).
Between June 5, 1996, and April 13, 1998, 310 women in labor with uncomplicated pregnancies were approached and 146 (47%) consented to continuous measurements of fetal arterial oxyhemoglobin saturation. Saturation was measured successfully in the fetuses of 129 women (88%); recording was not possible for technical reasons in the remaining 17 gravidas (12%). Our retrospective analysis was limited to the 129 women whose fetuses had fetal arterial oxyhemoglobin saturation measured. The mean age of those women was approximately 25 years, and 41 (32%) were nulliparous. Most (68%) were Hispanic, 24% were black, 6% were white, and 2% were Asian.
The fetal pulse oximeter sensor was placed at a mean (± standard deviation [SD]) cervical dilation of 5.3 ± 1.7 cm. The mean duration of fetal arterial oxyhemoglobin saturation monitoring was 3 hours and 12 minutes. A total of 135,345 10-second saturation averages were collected; the mean (± SD) number of 10-second averages per subject was 1049 ± 757, and the range of observations was 109–4012. Mean (± SD) fetal oxygen saturation was 51.7 ± 6.8% (95% confidence interval [CI] 38.3, 65.1) in the first stage of labor and 49.5 ± 9.6% (95% CI 30.6, 68.4) in the second stage. Those fetal arterial oxyhemoglobin saturation values were not significantly different from each other (P = .16).
A total of 69 fetuses (53%) had at least one 10-second averaged episode of arterial oxyhemoglobin saturation below 30% during labor. We analyzed the percentage of monitoring time during which saturation values less than 30% were registered: 126 of 129 fetuses had such values for less than 1% of their total monitoring times. Whereas fetal arterial oxyhemoglobin saturation values less than 30% were common (53% of cases), those episodes were typically transient during the entire course of saturation monitoring. We also analyzed intrapartum events in women with fetuses with saturation measurements less than 30% compared with women whose fetuses' saturation values were at least 30% (Table 1). The number of women who received oxytocin for stimulation of labor was significantly greater in the group whose fetuses had saturation values less than 30% (40 versus 21 fetuses with values at least 30%, P = .013). Fetal heart rate decelerations requiring interpretation by a physician, typically variable decelerations, occurred in 51 (74%) fetuses with values less than 30%, compared with 32 (53%) in those with values at least 30% (P = .017). As shown in Table 1, fetal arterial oxyhemoglobin saturation values below 30% were unrelated to presence of meconium-stained amniotic fluid, use of epidural analgesia, occurrence of prolonged second stage of labor, and route of delivery.
Indices of potential intrapartum fetal compromise were compared between the two groups (Table 2). There were no significant differences in the incidence of low 5-minute Apgar scores (3 or less), umbilical artery blood acidemia (pH less than 7.20), admission of term infants to the special care nursery, and cesarean delivery for nonreassuring FHR pattern. Six (10%) of those fetuses with arterial oxyhemoglobin saturation of at least 30% met at least one of the four criteria for fetal compromise, compared with eight (11%) of those with values less than 30% (P > .999).
Because of the frequent occurrence of fetal arterial oxyhemoglobin saturation values less than 30% and the failure of that threshold to identify fetal compromise during labor, we tested whether duration of fetal arterial oxyhemoglobin saturation below 30% correlated with poor fetal condition at birth. The incidence of the composite index of fetal compromise significantly increased as duration of fetal arterial oxyhemoglobin saturation below 30% increased from less than 1 minute to 9 minutes or more (P = .002 for trend analysis). Further analysis involved comparisons of the incidence of the composite index of fetal compromise for progressively longer durations of fetal arterial oxyhemoglobin saturation below 30% and the incidences for shorter durations (Table 3). There was a significant (P = .025) increase in the incidence of potential fetal compromise among fetuses with arterial oxyhemoglobin saturation below 30% for 2 minutes or longer. Using that 2-minute threshold, 11 (14%) of 78 fetuses with less than 2 minutes of arterial oxyhemoglobin saturation below 30% had potential fetal compromise, compared with seven (54%) of 13 with longer episodes (P < .001).
There were three significant findings in this study: fetal arterial oxyhemoglobin saturation values less than the predicted lower limit of 30% were common during normal labor, occurring in more than half of the fetuses we studied; transient fetal arterial oxyhemoglobin saturation values less than 30%, compared with those at least 30%, did not identify fetal compromise; and duration of fetal arterial oxyhemoglobin saturation below 30% might predict fetal compromise. In our series of 129 fetuses, as epochs of saturation below 30% increased from 1 to 9 minutes or more, fetal compromise incrementally increased in frequency. Compromise occurred when saturation was below 30% for 2 minutes or longer. Our results were consistent with the hypothesis that isolated fetal incidents, such as transient FHR decelerations, are common in labor and usually do not impair fetal oxygenation. Longer duration of fetal incidents in labor appears to be less common. It is the duration of an intrapartum fetal event that affects fetal oxygenation, rather than a moment-to-moment or beat-to-beat event such as an FHR deceleration.
Others have reported that the duration of fetal arterial oxyhemoglobin saturation values less than 30% might identify fetal compromise. Langer and colleagues8 averaged fetal arterial oxyhemoglobin saturation measurements during the final 10 minutes of labor in 41 women with abnormal FHR patterns. They found that fetal acidemia (pH no more than 7.15) was associated with an average fetal arterial oxyhemoglobin saturation value less than 30% during those final 10 minutes. Kühnert and colleagues9 studied 46 fetuses and concluded that fetal arterial oxyhemoglobin saturation values that averaged less than 30% for 10 minutes yielded a sensitivity of 81% and a specificity of 100% for predicting scalp pH below 7.20. In both of those studies, unlike our analysis, intervals less than 10 minutes were not examined.
Given the nonspecificity of electronic FHR monitoring in detecting fetal compromise, it is exciting that continuous fetal arterial oxyhemoglobin saturation measurement, with emphasis on the duration of hypoxia, could be an important adjunct for interpreting intrapartum FHR. However, our study sample was small and the outcomes of interest are too uncommon to extrapolate our results in 129 women without further studying a larger cohort. Also, we have not shown that a knowledge that the fetal arterial oxyhemoglobin saturation value is less than 30% for a specified period can be used to modify clinical outcomes. We believe that further clinical trials involving this promising new technology are warranted.
1. Dildy GA, Clark SL, Loucks CA. Intrapartum fetal pulse oximetry: Past, present, and future. Am J Obstet Gynecol 1996;175:1–9.
2. McNamara H, Chung DC, Lilford R, Johnson N. Do fetal pulse oximetry readings at delivery correlate with cord blood oxygenation and acidaemia? Br J Obstet Gynaecol 1992;99:735–8.
3. Luttkus A, Fengler TW, Friedmann W, Dudenhausen JW. Continuous monitoring of fetal oxygen saturation by pulse oximetry. Obstet Gynecol 1995;85:183–6.
4. Alshimmiri M, Bocking AD, Gagnon R, Natale R, Richardson BS. Prediction of umbilical artery base excess by intrapartum fetal oxygen saturation monitoring. Am J Obstet Gynecol 1997;177:775–9.
5. Nijland R, Jongsma HW, Nijhuis JG, van den Berg PP, Oeseburg B. Arterial oxygen saturation in relation to metabolic acidosis in fetal lambs. Am J Obstet Gynecol 1995;172:810–9.
6. Dildy GA, Thorp JA, Yeast JD, Clark SL. The relationship between oxygen saturation and pH in umbilical blood: Implications for intrapartum fetal oxygen saturation monitoring. Am J Obstet Gynecol 1996;175:682–7.
7. Goffinet F, Langer B, Carbonne B, Berkane N, Tardif D, Le Goueff F, et al. Multicenter study on the clinical value of fetal pulse oximetry. The French Study Group on Fetal Pulse Oximetry. Am J Obstet Gynecol 1997;177:1238–46.
8. Langer B, Carbonne B, Goffinet F, Le Goueff F, Berkane N, Laville M. Fetal pulse oximetry and fetal heart rate monitoring during stage II of labour. Eur J Obstet Gynecol Reprod Biol 1997; 72(Suppl):S57–61.
9. Kühnert M, Seelbach-Gölbel B, Butterwegge M. Predictive agreement between the fetal arterial oxygen saturation and fetal scalp pH: Results of the German multicenter study. Am J Obstet Gynecol 1998;178:330–5.