IN THE LIT

R. M. Pauli, M.D., Ph.D.

Stillbirths among offspring of male radiation workers at Sellafield nuclear reprocessing plant. Parker L, Pearce MS, Dickinson HO, Aitkin M, Craft AW. Lancet 354:1407-1414, 1999.

Occupational exposure to antineoplastic agents: self-reported miscarriages and stillbirths among nurses and pharmacists. Valanis B, Vollmer WM, Steele P. J Occup Environ Med 41:632-636.

These (and the articles that are reviewed below) are serendipitous pairings of publications that have appeared recently. This one is a natural one since antineoplastic agents and radiation have quite similar effects—most worrisome being the potential for mutagenic effects on germ cells. That, in turn, might lead to a variety of risks, including increased frequency of single gene disorders, of multifactorial birth defects and of lethal processes (resulting in miscarriage, stillbirth or neonatal death).

Workers at Sellafield experience a greater level of radiation exposure than anywhere else in the western world. This epidemiologic cohort study evaluated the pregnancy outcomes of the partners of (male) Sellafield workers with all others living in the surrounding county. Some of the statistical methods are tough sledding. Overall, Sellafield workers’ partners had only a very modest increased risk of stillbirth (relative risk of 1.14 with a 95% confidence interval between 0.82 and 1.88). However, a positive association was found between the probability for stillbirth (defined here as 28 weeks or greater gestation) and the total estimated exposure of the father. Such a dose-response relationship is suggestive of, but certainly does not prove, causality. Certainly it is biologically plausible that ionizing radiation exposure caused genetic damage to sperm precursors and that damage often resulted in lethality. Note, however, that the risk is very small; indeed, in a companion commentary (Inskip H: Stillbirth and paternal preconceptional radiation exposure. Lancet 354:1400-1401) it is pointed out that the increased risk today does not bring the total risk of stillbirth to the level faced just 10 years ago (since the total stillbirth rate has fallen).

Valanis et al. use a very different research design to try to assess if exposure to antineoplastics (which are also mutagenic) increases risk of pregnancy loss. This study is based on a questionnaire sent to nurses, pharmacists etc. who might tend to handle these agents. Measurements of frequency of intrauterine deaths are also dependent on self-reporting. So, one can anticipate marked inaccuracies of the latter and difficulty in really estimating dosage of exposure related to the former. Given all of these problems, I would fully expect that negative or uninterpretable data would result. The data, based on 7094 pregnancies, suggest a modest increased risk of stillbirth + miscarriage in women exposed shortly before or during (this implying a teratogenic effect) pregnancy as well as a statistically non-significant trend suggesting similar risks in partners of exposed men.

Overall, both studies suggest that there may be real but remarkably modest risks related to preconceptual exposure to ionizing radiation or occupational exposure to antineoplastic drugs. In both circumstances, in clinical practice it is probably appropriate to provide cautious reassurance.

Stillbirth as risk factor for depressions and anxiety in the subsequent pregnancy: cohort study. Hughes PM, Turton P, Evans CDH. BMJ 318:1712-1724, 1999.

The impact of perinatal loss on adjustment to subsequent pregnancy. Franche R-L, Mikail SF. Soc Sci Med 48:1613-1623, 1999.

Impact of perinatal loss on the subsequent pregnancy and self: women’s experience. Côté-Arsenault D, Mahlangu N. J Obstet Gynecol Neonat Nurs 28:274382, 1999.

I have listed these three articles in what is, in my opinion, their rank from best to worst. All address virtually identical questions related to the impact of previous pregnancy losses on emotional wellbeing in the next pregnancy. This is, of course, a very important issue, particularly if those insights might suggest specific helpful strategies for intervention.

Hughes et al. use a cohort method with matched groups of 60 women who experienced stillbirth (here 18 weeks gestation or more) and 60 controls (primigravida here). They used standard depression and anxiety scales to measure these during the third trimester of the subsequent pregnancy as well as at three times after the next delivery. Their major findings included the following. First, both depression and anxiety were greater in women who had experienced stillbirth (no surprise). Second, only among those who conceived again within 12 months of a loss, depression was increased 12 months after the next delivery. That is, women who conceive soon after stillbirth have more chance of prolonged depression than those who wait. This suggests that counseling couples to consider waiting, perhaps a year, before a next conception may be reasonable.

I liked Franche and Mikail’s article, if only (and pretty much only) because it looked at couples. The biggest problem with this (and one of the problems with the next) article is use of a strange definition of perinatal death. These authors define perinatal death as any death at any point in pregnancy or within 28 days of delivery. This is certainly nonstandard! Indeed, perinatal loss is usually defined as stillbirth beyond 20 weeks gestation plus death in the first 28 days of life. Why is this a problem? This article purports to talk about perinatal death when, in fact, the vast majority of couples assessed had experienced early miscarriage. Mixing three groups (miscarriage, stillbirth and neonatal death) implies that reactions to them are identical. That may or may not be so, but makes me suspicious of all of the data. These authors, too, used standard questionnaires with 31 pregnant women and 28 of their male partners. To sum, most pregnancies were characterized by mild depression, hypervigilance and considerable anxiety. One interesting finding is that only in women who think that they have a major role in maintaining fetal health was anxiety marked. This suggests an intervention strategy of some importance—emphasizing that most instances of loss are out of the control of the parents and are not their ‘fault’. The last interesting finding (yes this study was interesting if flawed in many ways) was the inverse relationship between dependency and level of depression (usually positively correlated). The authors suggest that “higher levels of dependency in both partners may be adaptive during the early stages of a pregnancy following a loss”—a strategy that some support groups have intuitively suggested.

Finally, the last of these three uses the same bizarre definition of perinatal loss. It provides a good review of some of the relevant literature. It explains its methods. And then it presents no quantitative results! Instead the authors wax on (and on). I think there are some insights here. I didn’t have the patience to wade through the verbiage when I realized it was solely descriptive and utterly interpretive.

Perinatal outcome in grand and great-grand multiparity: Effects of parity on obstetric risk factors. Babinszki A, Kerenyi T, Torok O, Grazi V, Lapinski RH, Berkowitz RI. Am J Obstet Gynecol 181:669-674, 1999.

Teenage pregnancies and risk of late fetal death and infant mortality. Olausson PO, Cnattingius S, Haglund B. Br J Obstet Gynecol 106:116-121, 1999.

Which is better—age and experience or youth? So far as fetal risks are concerned, quite clearly extreme youth of the mother yields far greater risks than does extreme experience (i.e. grand multiparity and great-grand multiparity).

Often, grand multiparity (pregnancy following 4 to 8 previous deliveries) and great-grand multiparity (more than 8 previous deliveries) are thought to carry with them special risks. This seems to be more based on fear of the relative infrequency of these events in our society than on much hard data. Babinszki et al. through a retrospective, cross-sectional study looked at a number of possible adverse outcomes in 314 grand multiparas, 133 great-grand multiparas and 2195 controls. Although (alas) they don’t show the data regarding intrauterine and neonatal death, they state that there were no significant differences among the three groups.

Olausson et al. looked more specifically at fetal death and infant mortality in the pregnancies of teenage girls using the Swedish Medical Birth Registry. They attempted to control for a number of variables (but could not, for example, control for diet, for gestation when prenatal care was begun etc.). They demonstrated a trend for increasing risk for late fetal death (here, »28 weeks gestation) with decreasing maternal age. More impressive is a risk for neonatal mortality in mothers between 13 years and 15 years of age. That latter risk appeared to be completely explained by an increased risk of preterm deliveries. Clearly, having a (usually) older and (perhaps) wiser great-grand multipara as a mother is much safer than if your mother-to-be is extremely young.

Influence of consanguinity and maternal education on risk of stillbirth and infant death in Norway, 1967-1993. Stoltenberg C, Magnus P, Terje Lie R, Daltvelt AK, Irgens LM. Am J Epidemiol 148:452-459, 1998.

Consanguinity and recurrence risk of stillbirth and infant death. Stoltenberg C, Magnus P, Skrondal A, Terje Lie R. Am J Pub Health 89:517-523, 1999.

Both of these studies take advantage of the registries that are unique to Scandinavian countries. Here, the Norwegian Medical Birth Registry was linked to other databases to derive the information needed for the analyses.

There is good theoretical reason to think that consanguinity (generally defined as relatedness of first cousins or greater between parents) might increase risk of stillbirth. Such consanguinity results in shared copies of recessive genes (shared because of identify by descent). So, the offspring of such unions would more likely be homozygous for harmful recessives, including lethals. How important this is has a broader implication — how important are genetic factors in instances of unexplained stillbirth?

First, there is a modest increased risk of stillbirth in infants of consanguineous parents, with an odds ratio of 1.3 (but confidence intervals of 1.0 to 1.6). This suggests that the contribution of lethal recessive genes is probably very small. Second, the recurrence risk for stillbirth in consanguineous unions is greater than in those that are not; this supports the notion of real, if modest, contribution of single gene, recessive causes to the etiology of stillbirth.

The second of these articles also provides, through its control data, one of the better resources for estimating recurrence risks following stillbirth. Here stillbirth is defined as 16 weeks or more and so is not absolutely comparable to our needs based on this country’s accepted definition (of 20 weeks or more). Nonetheless these Norwegian data do provide a good relative risk assessment: risk of stillbirth after it has already occurred once is just about 6 times as likely as in the general population. If stillbirth occurs in 1 in 125-150 deliveries (current best estimates in this country, correcting for problems of underreporting), then recurrence risk is around 4%—quite close to some previous estimates, including the estimate of 3%, which we have been using.