Authors' Reply

Abstract

be evidence that many of the fetuses have been chronically exposed to high levels of carbon monoxide3.It would have been instructive to know the maternal COHb levelsor better yet, plasma cotinine levels. In any case, if one accepts our contention that if the COHb levels in fetuses of nonsmoking mothers are indeed close to 1%, then the COHb levels due to smoking are not double, but at least six times higher than that of the nonsmokers. The full impact of smoking during pregnancy on the fetus remains uncertain but may be an important factor affecting neonatal outcome. AUTHORS' REPLY Sir,We are very grateful to Dr Vreman et aL's interest in the higher than expected control values of fetal COHb reported in our paper. In response to their points:1 . We note their confirmation of the good performance of our equipment at the high levels of COHb found in the blood of fetuses whose mothers smoke. 2. We do not believe our samples could have been contaminated at collection. 3. The correction for HbF assumed 90% of the blood was fetal haemoglobin and so should have been adequate. 4 . The study was undertaken in central London in an area with heavy traffic and high pollution levels. 5. We were concerned that smoking may have been under reported (which is why the quantity smoked was not included in the paper) but the patients knew the information as to whether they smoked or not was important for our research, and we believe this information to be accurate.We would like to add two more suggestions for the relatively high control levels. First, the data quoted in Vreman et al. 's letter suggests that newborn babies have lower COHb levels than fetuses and it is entirely possible that there are changes with gestational age, perhaps associated with the normal change in the proportion of HbF. Second, we did not allow for the effects of passive smoking. We agree with Vreman et al. that any elevation in the control data resulting from these potential problems would only have reduced the apparent statistical and clinical importance of our findings in smokers. The differences may therefore be even more significant than they appeared in the paper. The selective miscarriage of Down's syndrome from 10 weeks of pregnancy Sir, We would like to draw your readers' attention to an error in our paper (Vol 102, October 1995)', which was kindly brought to our attention by Jane Halliday and Lyn Watson of the Murdoch Institute in Melbourne. Table 1 shows the published estimates of the prevalence of Down's syndrome pregnancy at the time of chorion villus sampling (CVS), amniocentesis, and term (live births), together with our corrected figures. The reason for the correction is that the estimates of the prevalences of Down's syndrome at amniocentesis and at term were derived using maternal age at the time of CVS instead of at term, which we should have done to make the estimates comparable with the birth estimates at term. The revised figures indicate that between about 10 weeks of pregnancy and term an estimated 48% of Down's syndrome pregnancies ...