The article by Brennan and colleagues in this issue of the Archives describes a well-designed and carefully reported study of the linkages between exposure to maternal smoking during pregnancy and later crime. Using data from a large Scandinavian study of male births, the authors established that males whose mothers smoked during pregnancy had significantly higher rates of crime when studied into their middle 30s. These associations persisted when a range of confounding social, family, and maternal factors were taken into account. Results of the study also suggest that, although exposure to maternal smoking during pregnancy is related to life-course persistent offending, it is not related to adolescent-limited offending.

The study has several strengths, including the use of a large and representative birth cohort, the prospective assessment of the exposure variable and confounding factors, and the systematic collection of data on a well-defined outcome measure. Furthermore, the findings are consistent with those of a growing number of studies that demonstrate linkages between maternal smoking during pregnancy and later development of externalizing behaviors in middle childhood and adolescence. The present study extends these findings by showing that an association between smoking during pregnancy and antisocial behavior extends well into adulthood.

Taken together, the research evidence in this area leads to 3 general conclusions about the linkages between maternal smoking during pregnancy and later externalizing behavior problems.

First, there is generally consistent evidence to suggest that children who are exposed to maternal smoking during pregnancy are at increased risk of later externalizing behaviors that seem to extend over the life course.

Second, these associations seem to be resilient to control for confounding social, familial, and personal factors that are associated with maternal prenatal smoking.

Third, there is evidence of specificity of association. Results of a recent study found that, although exposure to maternal smoking during pregnancy was related to later antisocial behaviors, it was unrelated to other aspects of mental health and personal adjustment. Results of the study by Brennan et al also suggest that maternal prenatal smoking is specifically related to life-course persistent offending but not to adolescent-limited offending.

Collectively, these findings suggest that the evidence gathered in this area meets several criteria that are commonly used as a basis of causal inference in epidemiological studies. There is now evidence of a consistent and replicable association between maternal prenatal smoking and later antisocial behaviors. This association has been shown to be resilient to control for confounding and seems to involve a specific relationship between maternal prenatal smoking and antisocial behavior. However, although the accumulated evidence in this area is consistent with the hypothesis that smoking during pregnancy leads to an increased risk of later antisocial behavior among offspring, it would be unwise to draw strong or dogmatic conclusions on the basis of the current body of evidence. There are several important issues that need to be addressed before firm conclusions can be drawn. These issues include:

1. Identification of Underlying Mechanisms: An important issue to be addressed in establishing a cause-and-effect relationship between smoking during pregnancy and antisocial behavior concerns the underlying mechanisms by which persons who are exposed to maternal smoking during the antenatal period are placed at greater risk of later antisocial behavior. Several underlying mechanisms have been proposed, including fetal hypoxia, changes in serotonin uptake, changes in dopaminergic systems, and changes in DNA and RNA synthesis in the brain. However, these explanations remain speculative, and at present, to my knowledge, there is no well-established study of the underlying processes that may lead children who are exposed to maternal smoking during pregnancy to be at greater risk of later antisocial behaviors. These considerations suggest the need to embed the present research findings into a broader framework of theory and research about the relationships between prenatal events and later behavioral adjustment.

2. Genetic Confounding: Although existing studies have now controlled a relatively wide range of confounding social and environmental factors, there is an important source of confounding that has not been examined. It could be suggested that the linkages between maternal prenatal smoking and later adjustment reflect a genetic process in which the offspring of mothers who smoke during pregnancy are more likely to inherit genotypes that are associated with increased risk of later externalizing behaviors. There is some support for this view from previous research suggesting that a) cigarette smoking and antisocial behavior are highly comorbid {i.e., co-occur}, and b) externalizing behavior problems show high levels of heritability. One method of approaching this issue is through a twin design in which the linkages between smoking during pregnancy and later antisocial behavior are examined in monozygotic and dizygotic twins. Using such a design, it may be possible to examine the extent to which smoking during pregnancy contributes to later antisocial behavior when genetic effects on antisocial behaviors are taken into account.

The finding of consistent linkages between smoking during pregnancy and antisocial behaviors is intriguing, and these results invite the hypothesis that maternal smoking during pregnancy increases children's risk of later antisocial behavior. However, the uncertainties in these areas are such that it would be premature to conclude that maternal prenatal smoking can now be included among the established risk factors for later antisocial behaviors. There is further work to be conducted into underlying mechanisms and the possible confounding effects of genetic factors. Furthermore, the public health significance of the findings in this area is somewhat limited by the already well-established findings on the adverse effects of maternal smoking during pregnancy. These include increased risks of spontaneous abortion, reduced birth weight, compromised perinatal status, and lower intelligence. Given the ambiguities in the evidence noted above, perhaps the most prudent summation of research in this area is that maternal prenatal smoking may affect longer-term behavioral development, but considerable uncertainty still exists about the origins of the relationship. Irrespective of the ambiguities in this area, there is already abundant evidence to support the view that maternal smoking during pregnancy has undesirable consequences for children. Results of current research suggest that increased risk of later antisocial behavior may, in the future, become another outcome that will be added to the growing list of established adverse consequences of smoking during pregnancy.

EDITORIAL NOTE:
The results of this study suggest that maternal smoking during pregnancy increases the likelihood that offspring will engage in criminal activity in their adulthood as indicated by an arrest for a violent or nonviolent crime. Neither the authors' abstract, the excellent commentary by Dr. Fergusson, nor the newspaper account above mention that the relationship between maternal smoking and offspring antisocial behavior was only observed when there were birth complications. Also, unmentioned is the fact that birth complications were not related to maternal cigarette smoking in this study. We believe that this constitutes an important qualification to the study's findings and raises questions about the meaning of the research and its generalizability.

The study did not assess, and thus could not control for, maternal use of alcohol or illicit drugs during pregnancy or other factors that could potentially explain the observed relationship between maternal smoking and antisocial behavior of their offspring. Thus, future research is needed to rule out alternative explanations for the present findings.

A causal relationship between maternal smoking during pregnancy and criminal behavior of male offspring could have major public policy implications even if this effect only occurred when there were birth complications. Additional research is needed to identify underlying causal mechanisms and preventive interventions.

Over the years we have become concerned about the reporting of research to the general public. Sensationalistic reporting is reflected in the failure to present the latest study in the context of prior research and/or a failure to discuss adequately the qualifications or limitations of the research. We suspect that some researchers and journals may be complicit in this practice by oversimplifying research in press releases in order to increase press coverage. The potential danger of these practices is that the public may ignore all research.