Most of the focus of this blog is about the use of CTG monitoring during labour. A summary of the evidence about intrapartum CTG monitoring can be found here, here, here, and here. A summary of the evidence about admission CTGs can be found here. But what about using CTG monitoring during pregnancy? In this post I summarise the evidence which was systematically reviewed for Cochrane by Grivell et al. (2015).
Antenatal monitoring for women considered to be at risk
Research regarding the use of CTG monitoring in the antenatal period has focused exclusively on the use of CTG monitoring for women with pregnancy complications (such as pre-eclampsia) or women with other risk factors for poor perinatal outcome (such as reduced fetal movements or a suspected small baby). There have been six randomised controlled trials, including 2,105 women, which were conducted between 1982 and 1997. In two of these trials the comparison was between standard (clinician interpretation only) CTG and computer assisted CTG interpretation (Bracero et al., 1999; Steyn & Odendaal, 1997), three compared CTG monitoring with the findings revealed to the treating team against CTG monitoring with the findings concealed (Brown et al., 1982; Flynn et al., 1982; Kidd et al., 1985), and the remaining trial compared CTG monitoring with no CTG monitoring (Lumley et al., 1983). It is important to note that the comparator here was not intermittent auscultation, and the trials were conducted during a period where there was less use of ultrasound assessment of fetal wellbeing.
Placing the research into context
By 1982, when the first of the trials regarding antenatal CTG monitoring was published, five RCTs had been conducted examining the use of intrapartum CTG monitoring (Haverkamp et al., 1979; Haverkamp et al., 1976; Kelso et al., 1978; Renou et al., 1976; Wood et al., 1981), with four showing no benefit to the fetus. Despite this mounting evidence that CTG monitoring was not able to achieve the goals it seemed to promise, clinicians were expanding the use of CTG monitoring into the antenatal period. In 1983, Lumley et al., said that antenatal CTG monitoring “has become an established procedure for the management of the high-risk pregnant woman and has been extended to the entire obstetric population in a few centres” (p. 1018). There are many such examples of the triumph of hope over reality in relation to CTG monitoring research.
After a flurry of research activity over the three-year period of 1982 to 1985, the research comparing antenatal CTG monitoring with no monitoring (or where the results were blinded to clinicians) stopped and was not picked up again. The remaining two papers in the 90s didn’t aim to demonstrate whether or not antenatal CTG monitoring was of benefit, but asked whether adding computer interpretation of the trace to CTG monitoring would improve outcomes.
How might antenatal CTG monitoring work?
Before examining the details of the research findings, it is worth stopping for a moment and asking about the mechanism by which antenatal CTG monitoring might be able to modify perinatal outcomes. Knowing this helps to understand what outcome measures to look at in the research. For example, a finding that women with abnormal fetal heart rate patterns experienced a poor perinatal outcome more often than women with normal patterns, while interesting, does not demonstrate that knowing that the fetal heart rate pattern is abnormal leads to a change in management, nor that changing her care ultimately leads to a better outcome.
The use of antenatal CTG monitoring is based on an assumption that there is a period of time preceding stillbirth or neurological injury in which the fetal heart rate pattern will be abnormal. By identifying fetuses with this particular abnormal pattern, prompt birth is recommended to prevent stillbirth or long-term neurological damage. This assumes that there is a predictable, meaningful, pattern of abnormality that reliably proceeds damage to the fetus (which is yet to be demonstrated in research).
Antenatal CTG monitoring with the findings revealed vs CTG monitoring with the findings concealed or no CTG
All the four studies that compared CTG use with no CTG, or concealed the findings from clinicians were considered to be of poor quality, and were small, with a total of 1,627 women in the studies. There were no statistically significant differences in the caesarean section rate, gestational age at birth, the use of induction of labour, low Apgar scores at five minutes of age, or in the rate of admission to a neonatal nursery. The absence of a difference in caesarean sections and inductions of labour might be because clinicians didn’t act on the additional information available from the CTG, or (as is highly likely) the size of the research was simply to small to detect a difference if there actually was one. There was also no significant difference in the rate of seizures (occurring in 1 of 144 babies with CTG monitoring antenatally and 2 of 156 babies who were not exposed to CTG monitoring). The incidence of long term neurological outcomes were not measured, nor were women’s experiences with antenatal CTG monitoring.
The perinatal mortality rate for babies who had received antenatal CTG monitoring was 23 per 1000 births and was only 11 per 1000 births when CTG monitoring was NOT used. CTG monitoring was therefore associated with a doubling of the perinatal mortality rate (RR 2.05, 95% CI 0.95 – 4.42), however the small number of babies in the population means that the research was underpowered and therefore not able to demonstrate a statistically significant difference, even if CTG use really does double the rate of deaths. When only babies without significant congenital abnormalities were considered, the mortality rate was two and a half times lower when CTG monitoring was NOT used (6 per 1000 compared to 17 per 1000 births when CTG monitoring was used). This once again came close to, but did not achieve statistical significance (RR 2.46, 95% CI 0.96 – 6.3).
Computerised vs standard CTG monitoring
It is worth first pointing out that the two studies which examined computer assisted CTG interpretation were both done in the 1990s, when the technology available was rather primitive compared to what is now commercially available for intrapartum use (Bracero et al., 1999; Steyn & Odendaal, 1997). In addition, while the quality of the studies was better (rated as moderate), the size of the studies was small, with only 469 women. As a consequence, caution should be taken in generalising these findings to current practice.
No significant differences were found in the caesarean section rate, low Apgar scores at five minutes of age, gestational age at birth, or the length of stay in the nursery. Again, no data were collected about long term neurological outcomes, or women’s experiences of monitoring. The perinatal mortality rate was 9 per 1000 births in babies who had been monitored with the computerised CTG and 43 per 1000 for standard CTG monitoring, a statistically significant difference (RR 0.2, 95% CI 0.04 – 0.88). When babies with congenital abnormalities were excluded, the perinatal mortality rate was 4 per 1000 with computerised CTG monitoring and 26 per 1000 births for standard CTG monitoring, which was no longer statistically significant (RR 0.23, 95% CI 0.04 – 1.29).
Rather than providing firm support for the use of antenatal CTG monitoring for women who are considered to be a high risk for a poor perinatal outcome, the research suggests that CTG monitoring may possibly cause harm. I find it curious that in the face of evidence suggesting a trend towards a higher mortality rate, further research was not conducted, and that antenatal CTG monitoring has continued to be used. The assumption that CTG monitoring works is so strong, it seems that researchers may have been anxious to avoid pursuing further research which might have definitively undermined this faith.
I am equally curious as to why computerised CTG monitoring for antenatal use was not mandated on the back of the evidence of a reduction in perinatal mortality rates, or at least provided the impetus for a large, multi-centre trial, adequately powered to address whether the findings from these small studies could be relied upon in practice. Given the obstetric love of all things technological, I am at a loss to explain why this has not been revisited in the past twenty years. The question of whether computer assisted CTG monitoring is associated with a reduction in perinatal mortality compared to no CTG monitoring also remains unanswered.
There is one particular corner of clinical practice where current antenatal fetal monitoring practices make me more uncomfortable than any other: the use of CTG monitoring for women who present with concerns about reduced fetal movements. Women with this concern were included within the populations studied in the RCTs summarised here. At best, CTG use in this population is useless at preventing harm, and at worst, might actually lead to higher rates of death. The mechanisms for this are not currently known. Perhaps women and their clinicians place misguided faith in the CTG output rather than taking more timely action? Any policies regarding the management of women with reduced fetal movements must require a comprehensive assessment of the woman, including ultrasound scanning. “I’ll just pop on a quick CTG and see if the baby is happy” is not an evidence-based assessment approach, and may do harm.
Burying our collective heads in the CTG sand while pretending that CTG monitoring is an effective tool is preventing us from doing the research we need to address the really important questions of how best to care for women during the antenatal period and during labour to prevent the death or damage of their baby from low oxygen levels.
Bracero, L. A., Morgan, S., & Byrne, D. W. (1999). Comparison of visual and computerized interpretation of nonstress test results in a randomized controlled trial. American Journal of Obstetrics and Gynecology, 181(5 Pt 1), 1254-1258.
Brown, V. A., Sawers, R. S., Parsons, R. J., Duncan, S. L. B., & Cooke, I. D. (1982). The value of antenatal cardiotocography in the management of high-risk pregnancy: a randomized controlled trial. BJOG: An International Journal of Obstetrics and Gynaecology, 89(9), 716-722.
Flynn, A. M., Kelly, J., Mansfield, H., Needham, P., O'Conor, M., & Viegas, O. (1982, Jun). A randomized controlled trial of non-stress antepartum cardiotocography. BJOG: An International Journal of Obstetrics and Gynaecology, 89(6), 427-433.
Grivell, R. M., Alfirevic, Z., Gyte, G. M. L., & Devane, D. (2015). Antenatal cardiotocography for fetal assessment. Cochrane Database of Systematic Reviews, 9, CD007863.
Haverkamp, A. D., Orleans, M., Langendoerfer, S., McFee, J. G., Murphy, J., & Thompson, H. E. (1979). A controlled trial of the differential effects of intrapartum fetal monitoring. American Journal of Obstetrics and Gynecology, 134(4), 399-412.
Haverkamp, A. D., Thompson, H. E., McFee, J. G., & Cetrulo, C. (1976, Jun 01). The evaluation of continuous fetal heart rate monitoring in high-risk pregnancy. American Journal of Obstetrics and Gynecology, 125(3), 310-320.
Kelso, I. M., Parsons, R. J., Lawrence, G. F., Arora, S. S., Edmonds, D. K., & Cooke, I. D. (1978). An assessment of continuous fetal heart rate monitoring in labor. A randomized trial. American Journal of Obstetrics and Gynecology, 131(5), 526-532.
Kidd, L. C., Patel, N. B., & Smith, R. (1985, Nov). Non-stress antenatal cardiotocography–a prospective randomized clinical trial. BJOG: An International Journal of Obstetrics and Gynaecology, 92(11), 1156-1159.
Lumley, J., Lester, A., Anderson, I., Renou, P., & Wood, C. (1983). A randomized trial of weekly cardiotocography in high-risk obstetric patients. BJOG: An International Journal of Obstetrics and Gynaecology, 90(11), 1018-1026.
Renou, P., Chang, A., Anderson, I., & Wood, C. (1976). Controlled trial of fetal intensive care. American Journal of Obstetrics and Gynecology, 126(4), 470-476.
Steyn, D. W., & Odendaal, H. J. (1997). Routine or computerized cardiotocography in severe preeclampsia? A randomized controlled tiral. Journal of Maternal Fetal Investigation, 7, 166-171.
Wood, C., Renou, P., Oats, J. J. N., Farrell, E.-M. E., Beischer, N., & Anderson, I. (1981). A controlled trial of fetal heart rate monitoring in a low-risk obstetric population. American Journal of Obstetrics and Gynecology, 141(5), 527-534.