The authors of a research study carried out in France have echoed existing concerns that trying to identify small babies during pregnancy may do more harm than good.
In a freely available paper published in the BJOG, Monier et al (2014) begin by describing the current problem: although we know that a few babies are on the small side for their gestational age (SGA), and are thought to be at increased risk as a result of this, there is nothing (that we know of that has been adequately researched*) that we can do about this during pregnancy. Except make women worry and feel inadequate, which sadly is achieved all too effectively with the current approach taken in mainstream practice. (*I’ll come back to this below).
At the present time, the best that we can offer is to try and identify these babies in utero, by means of screening tests which seek to measure them. We measure women’s abdomens with our hands and our tape measures, and in some areas with ultrasound, and we map what we think we feel or find or see against charts showing the average size of babies at various stages of pregnancy. If these screening tests cause us to we think we have found a baby who is ‘small for gestational age’, then the only thing that the maternity services can offer is to deliver the baby early. Which means intervention that carries more risks. Such an offer also sends a very clear message from the medical profession to the woman that she isn’t good enough, which can have really negative consequences, but a full discussion of that important issue is perhaps best saved for another day.
Aware that these screening tests are “a central component of antenatal care” (2) but that previous research studies into this area are small, old and/or not inclusive of babies who were suspected of being SGA but who turned out to be of ‘normal’ birthweight, Monier et al (2014) undertook a huge analysis of data from the French National Perinatal Survey. This data set included all live and stillborn babies who were born after 22 weeks of gestation with a birthweight of at least 500g in all maternity units in France over a one week period. The total number of babies was 14,100. The authors acknowledge that the use of retrospective data always means there is a risk that the numbers are slightly out because researchers have to rely on medical records and can’t always know whether some things weren’t written down, but they make the point that their results are so significant that even a ten per cent margin of error wouldn’t make that much difference.
Like some of the previous studies which showed that only between 10% – 36% of babies with a birthweight under the 10th percentile are detected during pregnancy, Monier et al (2014) found that only 21.5% of the babies who turned out to have SGA were suspected of having this during their mother’s pregnancy. This confirms that we are still not very good at detecting SGA, which is one problem.
Of just as much concern, though, is that half of the babies who were thought to be small – and whose mothers were told they were small, and whose mothers had extra screening and sometimes intervention (including induction of labour) because they were thought to be small – turned out not to have been small at all. So of women who were told that their baby was too small, only half of them actually had a baby that was small. There was just as much chance that the assessment was wrong as that it was right.
These figures are quite shocking. The rate of detecting genuinely small babies is low. The number of babies who are inaccurately identified as being too small (which triggers further screening, interventions and worry to their mums) is high. Monier et al (2014) conclude from their results that “the effectiveness of ultrasound and symphysis–fundal height measurement for the detection of FGR is poor.”
Surely we need a complete rethink of our approach in this area?
Which returns me to the tangent I mentioned to near the beginning of this post. I said that there is nothing that we know of that has been adequately researched that we can do about babies’ growth during pregnancy, and that is indeed the mainstream medical viewpoint. But if you talk to some midwives, they have a very different perspective. They believe, often from years of experience (but no, this has not been adequately researched), that how we feed our babies matters. Mostly they mean ‘feed’ in relation to nutrition, which itself is all but ignored in many areas, but some will talk about the notion of feeding in the widest sense of the word, and in relation to a woman’s need for sunlight, positivity, love, heart-centredness and all sorts of other good stuff that doesn’t get measured in randomised controlled trials, like feeding ones’ soul by knitting or gardening. And a lot of people think that’s all a bit cosmic. But, given the shocking figures that Monier et al (2014) have found, the fact that we don’t have a good way of pinpointing more than one in five of these babies and the volume of worry that is being generated in women who are needlessly and erroneously told that their baby is too small, maybe a completely new approach wouldn’t be such a mad, bad, cosmic idea after all?
Objective: To assess the proportion of small for gestational age (SGA) and normal birthweight infants suspected of fetal growth restriction (FGR) during pregnancy, and to investigate obstetric and neonatal outcomes by suspicion of FGR and SGA status at birth.
Design: Population-based study.
Setting: All French maternity units in 2010.
Population: Representative sample of singleton births (n = 14100)
Methods: We compared SGA infants with a birthweight of less than the 10th percentile suspected of FGR, defined as mention of FGR in medical charts (true positives), non-SGA infants suspected of FGR (false positives), SGA infants without suspicion of FGR (false negatives) and non-SGA infants without suspicion of FGR (true negatives). Multivariable analyses were adjusted for maternal and neonatal characteristics hypothesised to affect closer surveillance for FGR and our outcomes.
Main outcome measures: Obstetric management (caesarean, provider-initiated preterm and early term delivery) and neonatal outcomes (late fetal death, preterm birth, Apgar score, resuscitation at birth).
Results: 21.7% of SGA infants (n = 265) and 2.1% of non-SGA infants (n = 271) were suspected of FGR during pregnancy. Compared with true negatives, provider-initiated preterm deliveries were higher for true and false positives (adjusted risk ratio [aRR], 6.1 [95% CI, 3.8–9.8] and 4.6 [95% CI, 3.2–6.7]), but not for false negatives (aRR, 1.1 [95% CI, 0.6–1.9]). Neonatal outcomes were not better for SGA infants if FGR was suspected.
Conclusion: Antenatal suspicion of FGR among SGA infants was low and one-half of infants suspected of FGR were not SGA. The increased risk of provider-initiated delivery observed in non-SGA infants suspected of FGR raises concerns about the iatrogenic consequences of screening.