Adaptation of pulse oximetry for fetal monitoring during labour
Application of pulse oximetry to intrapartum monitoring was investigated in 105 women. No adequate reading could be obtained in 44 cases. Two major sources of artifact, related to probe apposition and signal processing, were identified and excluded. The average arterial oxygen saturation from the fetal scalp was 82% (SD 6%), which is higher than has been inferred from pO2 levels. Readings below 60% or the development of an unstable baseline suggest the presence of substantial
hypoxia and acidosis.
and R wave of ECG) are not synchronous. We can therefore ensure that even at these low signal amplitudes, only systolic, arterial oxygen saturation values are recorded.2 Pulse oximeter sensors can be used in transmission mode (eg, on the finger) or as a reflectance probe, with the photodiode and light-emitting diode (LED) placed next to each other (eg, on the forehead). Inappropriate apposition of sensors may result in interference from ambient light and most oximeters can display a warning if this occurs. However, we have observed that in the absence of ambient light, as in the birth canal, inadequate sensor-to-skin contact is a source of a different type of artifact: a gap between probe and skin may give a plethysmographic waveform of normal appearance (fig 1), but a proportion of light from the LED will be reflected straight back from the skin surface.
Introduction has revolutionised monitoring in intensive anaesthesia, care, and neonatal care. The principal aim of fetal monitoring during labour is to detect hypoxia and its effects, and pulse oximetry could have an important role here also. However, this application has so far been hampered by the limited access to the fetus. We aimed to identify and overcome these difficulties and to investigate the possible role of pulse oximetry in intrapartum Pulse
monitoring. Methods Oximetry utilises the fact that the light absorption spectra of and reduced haemoglobins differ, a characteristic of both fetal and adult haemoglobins.l A pulse oximeter uses two
reference wavelengths, one in the red and one in the near-infrared part of the spectrum (eg, 660 nm and 940 nm). At both wavelengths, absorption occurs in skin, subcutaneous fat, and connective tissue (the non-pulsatile component) as well as in the pulsatile vascular bed. The pulse oximeter is mainly concerned with the pulsatile, alternating component of the signal; the ratio of non-absorbed pulsatile/non-pulsatile components is calculated for both wavelengths and the non-pulsatile values are allowed to cancel each other out. The ratio of the two pulsatile components (red/infrared) is then used through an algorithm to calculate the percentage of haemoglobin that is saturated; if this ratio is calculated at the peaks of the pulsatile signal, the pulse oximeter can be said to indicate oxygen saturation levels of arterial blood (Sa02). Because the pulse oximeter uses ratios rather than absolute values, saturation measurements can be obtained even if the amplitude of the signal is low-for example, if the source is the fetus during labour. The autogain of most pulse oximeters magnifies the signal so that a waveform can be obtained, but the non-pulsatile component of the signal or "background noise" is therefore also amplified; if an accurate SaOzis to be derived from the ratio of amplified absorbances against a background of increased noise, it is even more important that the peaks of the plethysmographic waveform are precisely identified. We used the R wave of the fetal electrocardiogram (ECG), which is already obtained for conventional electronic heart rate monitoring, together with the pulsatile signal obtained with the photosensor. We had the synchronisation programme of our oximeter (CSI 504US, Criticare Systems Inc, Wankesha, Wisconsin, USA) modified so that the pulsatile signal is enhanced and averaged to exclude motion artifact, but also so that oxygen saturation measurements are rejected if the two signals (peak of
Fig 1-Effect of inadequate sensor-to-skin contact. Upper traces: incomplete apposition of reflectance probe on adult finger, in absence of ambient light: "normal" pulsatile waveforms are obtained but the saturations are falsely low. Lower panel: Increasing gap between sensors and skin is mirrored by artifactual drip in saturation readings (99% to 57%); saturation increases to normal with return of sensors to full contact.
Of the two wavelengths at which the oximeter operates, infrared permeates tissue better (ie, is less absorbed) and is similarly less absorbed by the air-skin interface. This affects the red/infrared ratio of the pulsating and non-pulsating components in a different way,
resulting in an increase in the red/infrared ratio which is translated by the oximeter’s algorithm as a lower saturation value. The magnitude of this error is a function of the distance between probe and skin (fig 1). We have quantified this effect in vitro.3 Thus good apposition of the probe is very important, and we have aimed to achieve this with a double clip design.2 Modified oximeter reflectance sensors (Criticare Systems Inc) are mounted between a ADDRESS:
Department of Obstetrics and Gynaecology, Queen’s Medical Centre, Nottingham NG7 2UH, UK (J. O. Gardosi, MRCOG, C. M. Schram, MRCOG, E. M Symonds, FRCOG). Correspondence to Mr Jason O. Gardosi.
The oxygen saturation trend remained steady during normal labours and most baselines ranged from about 70% to 90% (mean 82% [SD 6%]). Baseline readings from the breech were lower (50-65%; fig 2A). Physiological contractions showed no effect or small dips in Sa02 (fig 2B) but uterine hyperstimulation due to excessive oxytocin infusion caused an acute and reversible drop in saturation readings. Fetuses showing cardiotocographic abnormalities such as variable decelerations, but who had normal scalp blood-gas analysis and normal Apgar scores on delivery, also maintained a stable Sa02 baseline. However, in 4 fetuses evidence of metabolic acidosis (pH below 72, base deficit more than 10) developed; all 4 showed an unstable Sa02 baseline and falls of the Sa02 trend to 60% or below (fig 3), which preceded the diagnosis made on repeated scalp blood
sampling. Discussion Fig 2-Examples of cardiotocographs (top=fetal heart, lower=contractions) with Sa02 displayed on second ("toco") channel. A: probe applied on breech. Oximeter switched on at arrow. B: variable decelerations, fetal scalp blood sample normal (pH 7,32, base deficit 5).
The problems and limitations of pulse oximetry need to be recognised in this new application, as they have been in others.4-6 Attention to possible sources of artifact may explain why our findings differ from previous results.’°8 Our baseline oxygen saturations are higher, and accord with pulse oximetry measurements during fetal surgery9 as well as baseline recordings from our ongoing fetal lamb studies. Although the p02 in the fetal circulation is low, the left shift of the fetal haemoglobin dissociation curve results in avid O2 binding and high saturation levels. The degree of the shift is proportional to the fetal/adult haemoglobin ratio, which can vary substantially among fetuses this variation may explain some of the range in baseline Sa02 values we observed. The finding of lower Sa02 levels on the breech than on the scalp is consistent with the physiology of the fetal circulation, since blood below the ductus arteriosus is less well oxygenated. One source of artifact is venous and oedematous pulsations. Caput succedaneum on the neonatal scalp may cause a fall in oximeter readings," which is due to a false (non-arterial) pulse rate that can be recognised and excluded by synchronisation of the electronic (ECG) with the optical
(plethysmograph) signals.2,12 Fig 3-Baseline Sa02 during labour. Preterm fetus= 32 wk.
pair of Copeland-type arcuate needles (Surgicraft Ltd, Redditch, Worcs, UK), which are at the same time used as scalp electrodes for conventional fetal heart rate monitoring. The probe can be applied any time after the cervix is more than 2 cm dilated and the membranes are ruptured and it can be left on the fetal scalp until delivery. We are thus able to obtain a continuous recording of fetal Sa02 throughout labour which can be displayed directly onto the cardiotocograph strip (on the second "toco" channel; fig 2). The probe was applied in a series of unselected women in labour at St Mary’s Hospital, London, and Queen’s Medical Centre, Nottingham, with approval of the relevant ethics committees and maternal consent. The study was observational only and labour was managed according to usual practice, which included fetal scalp blood sampling in cases of suspected fetal distress.
oximetry probe were used in 105 In 44 (42%) cases no adequate reading was obtained after a maximum of two probe placements; this failure was usually due to the fetus’ having thick, curly hair. In the successful cases, oximetry monitoring lasted for 25 min to 85 h. Prototypes
Another artifact may arise from inadequate probe contact. Previous studies used tissue pressure or suction probes,8 but these cannot guarantee good apposition in the birth canal. Johnson and colleagues’8 inadvertent finding of Sa02 increases during contractions suggests poor sensor application (see fig 3 of their paper): during each contraction the fetal head would be pressed into better contact with a loose probe, increasing temporarily the saturation reading to what should be the true baseline value. Our double-clip probe was designed to ensure contact, but failed when the fetus had thick hair. We are now developing a non-invasive probe which is pressed by a balloon into good contact with the side of the fetal head, thus avoiding hair and caput
altogether. We see fetal oximetry as a potential adjunct to heart rate monitoring which can provide a continuous assessment of relative changes-ie, trends. It may have several uses in intrapartum monitoring. Fetal oximetry could confirm fetal wellbeing non-invasively. A stable Sa02 baseline would reassure observers that a questionable fetal heart trace represented only physiological stress during labour. Fetal scalp blood sampling, an invasive procedure which needs to be repeated to give an indication of trend, could then become unnecessary in most instances. A second use is to assess fetal
2. Gardosi J, Carter M, Becket T. Continuous intrapartum fetal oxygen saturation. Lancet 1989; ii: 692-93.
low baseline saturation at the beginning of labour, which we have seen in several cases of growth retardation (fig 2), could suggest diminished placental transfer and reserve. The technique could also be used to recognise fetal distress: a fall in the Sa02 trend during labour, which was observed during developing metabolic acidosis, may be due to movement of the haemoglobin dissociation curve to the right (Bohr shift); this on its own would reduce Sa02 even if P02 remained constant. These hypotheses will need to be tested in large series. Our aim was to assure the quality of our data and to exclude possible sources of artifact, to give this new fetal monitoring technique every chance to fulfil its exciting reserve: a
3. Damianou D, Crowe JA, Schram CMH, Gardosi JO. Artefacts in reflection pulse oximetry readings due to variations in probe to skin surface separation. IEE Abstr 1991; 89: 3. 4. Tremper KK, Barker SJ. Pulse oximetry. Anesthesiology 1989; 70: 98-108. 5. Southall DP, Bignall S, Stebbens VA, Alexander JR, Rivers RPA, Lissauer T. Pulse oximeter and transcutaneous arterial oxygen measurements in neonatal and paediatric intensive care. Arch Dis Child 1987; 62: 882-88. 6. Editorial. The trust in pulse oximeters. Lancet 1990; 335: 1130-31. 7. Peat S, Booker M, Lanigan C. Ponte J. Continuous intrapartum measurement of fetal oxygen saturation. Lancet 1988; ii: 213. 8. Johnson N, Johnson VA, Fisher J, Jobbings B, Bannister J, Lilford RJ. Fetal monitoring with pulse oximetry. Br J Obstet Gynaecol 1991; 98: 36-41. 9. Rosen M. Fetal surveillance during fetal surgery. Presented at 1st International Symposium of Intrapartum Surveillance, Nottingham, October, 1990. 10. Andrews BF, Willet GP. Fetal hemoglobin concentration in the newborn. Am J Obstet Gynecol 1965; 91: 85-88. 11. Johnson N, Johnson VA, Bannister J, Lilford R. The effect of caput succedaneum on oxygen saturation measurements. Br J Obstet
promise. This work was supported by Action Research, Nottingham Health Authority, Parkside Health Authority, Criticare Systems Inc, and Surgicraft Ltd.
REFERENCES 1. Harris AP, Sendak MJ, Donham RT, Thomas M, Duncan D. Absorption characteristics of human fetal hemoglobin at wavelengths used in pulse oximetry. J Clin Monit 1988; 4: 175-77.
Gynaecol 1990; 9: 493-98. 12. Schram CMH, Gardosi J. The effect of caput succedaneum on oxygen saturation measurements. Br J Obstet Gynaecol 1991; 98: 113-14.
PUBLIC HEALTH Africa’s 3 million blind people?
Who will operate
About half the 6 million blind people in sub-Saharan Africa have surgically curable cataract. The available manpower and resources can only provide services for less than 10% of the new blind cataract patients each year, and little is being done for the estimated 3 million "cataract backlog". A serious limiting factor to the development of prevention of blindness programmes is lack of trained manpower. Despite an increase in the number of ophthalmologists trained in cataract surgery (which varies greatly from country to country), this number is not keeping pace with increased demand for eye-care services, especially in large rural popultions. Initiatives that will help to overcome this dilemma are specific post-graduate courses in community ophthalmology in Africa, plans to develop a one-year diploma in ophthalmology course for English-speaking West African countries, and a proposal to upgrade a similar course in Zimbabwe. Additionally there is a need for the training of more ophthalmic assistants, cataract surgeons, and nurses in the diagnosis and management of common ophthalmic disorders. Experienced expatriate ophthalmologists also have an important role in the teaching of doctors and ophthalmic assistants how to select patients and carry out successful inexpensive cataract surgery with appropriate technology and limited facilities.
Introduction There are 27-35 million blind people world wide:1 6 million live in sub-Saharan Africa, of whom more than half have surgically curable cataract. The average ophthalmologist in Africa carries out fewer than 300 cataract operations a year on about 100 blind people. Therefore, 50 000 blind cataract patients are being operated on per year in sub-Saharan Africa, but the estimated "cataract backlog" is in excess of 3 million and there is an annual incidence of at least 500 000 newly blind people due to this disorder. Thus, the available manpower and resources are providing services for less than 10 % of new blind cataract patients, and there is still the problem of the existing cataract blind. A serious limiting factor to the development of prevention of blindness programmes in sub-Saharan Africa is the lack of trained manpower. In this article I will review the present situation about available personnel for providing cataract surgical services in Africa so that I can try to answer the question, "Who will operate on Africa’s 3 million curably blind people?"
Ophthalmologists The number of available ophthalmologists varies greatly from one region of the world to another, with about 1
ophthalmologist per 20 000 population in North America and western Europe, 1 per 35 000 in Latin America, and ADDRESSES International Centre for Eye Health, 27-29 Cayton Street, London, EC1V9EJ, UK (Allen Foster, FRCS).