Moderate Drinking : How Risky for Children Exposed In Utero?
Once again, a new suite of studies of the effects of moderate drinking on children’s learning capacities (1) that does not find a connection to prenatal alcohol consumption under 9 drinks a week has raised concerns among the Fetal Alcohol Spectrum Disorders (FASD) research and advocacy community as news and media outlets have widely reported these results. Studies of 870 children whose mothers reported low (1-4 drinks per week) or moderate drinking (5- 8 glasses weekly) found no significant relationship of that drinking to executive function (2), selective and sustained attention (3), or IQ (4) among 5 year old children. The moderate drinking group was compared to a matched group of 758, both selected from the Danish National Birth Cohort, a very large sample of 101,042 pregnancies (5).
Caution in interpreting these results has come from researchers who note, among other things, that age 5 may still be too young to find significant neurodevelopmental learning outcomes with this level of exposure (6). While there may have been other conflicting findings in the literature, several observers including The National Organization of Fetal Alcohol Syndrome (NOFAS) have noted that many other studies have found significant negative outcomes including spontaneous abortions or miscarriages when mothers reported drinking at a moderate level during pregnancy (7). Like NOFAS, we note that significant risk relationships between moderate drinking in early pregnancy and spontaneous abortion were reported in a recent (2011) secondary analysis of data from the full Danish National Birth Cohort (8) – the same database that has generated the lack of significant learning findings with 5-year-old children whose mothers drank low or moderate amounts of alcohol while pregnant.
Ah ha! How can the significant results of moderate drinking and spontaneous abortions be reconciled against the lack of association of moderate drinking with learning outcomes arising from the same sample? One wonders if the existence of these prenatal deaths might have been disregarded in the study of prenatal alcohol survivors. Is there a way to calculate cumulative risk in order to provide a more comprehensive and accurate picture of the effects of moderate drinking?
More questions come to mind. How different is the approach taken in the sampling process (5) from the full Danish birth cohort compared to a matched sample design? How are missing data treated? Is the sample skewed? Does the sampling process lead to more or less representativeness than the original full sample? Is there weighted data involved in the original sampling plan?
There are traditional statistical ways to deal with missing data. Imputation (9) is a statistical method used to account for missing data in a longitudinal study when there is non-response for other reasons. The typical methods for imputation assume that data that is missing is missing at random. With evidence that moderate drinking in the Danish National Birth Cohort was a risk factor for miscarriage, it seems likely that this is not a random outcome. Missingness (9) that is not at random or “nonignorable” is apparently typically biased. The creation of a fair plan to consider the missing subjects is vital with this and other studies. Any known prior findings with moderate drinking in any particular data base should be communicated, along with the results to the public. We will leave this conundrum in the hands of the able researchers – both those who likely did not have access to this information at the formation of the study and others whose research is also subject to these methodological problems.
In fact, the researchers worked very carefully to create a model plan for analysis and implement it carefully. They do note however that, “we used data from a homogenous, generally middle-class population with access to comprehensive healthcare free of charge for the individual (thereby reducing the potential for confounding) (5). What may have been gained in simplicity may have been lost in relevance and representativeness. Are the data being reported in newscasts as representative only of a middle-class population with access to free health care? Skewness may have been naturally occurring as higher SES could itself be a confounding variable associated with fewer miscarriages with that population. In any case, such a skewed sample is not representative of the whole sample of pregnant women who did and did not report drinking moderate amounts during pregnancy for this major longitudinal study. Without a full discussion of the methodology that we hope to see soon, these questions may go unanswered.
Unless consumers of research ask these kinds of questions, we may be promoting damage to the unborn by sanctioning prenatal alcohol at levels that are dangerous. For now, remember this, from a study of 92,719 participants in the Danish National Birth Cohort, up to 16 weeks into pregnancy, the researchers’ conclusions were, “Even low amounts of alcohol consumption during early pregnancy increased the risk of spontaneous abortion substantially. The results indicate that the fetus is particularly susceptible to alcohol exposure early in pregnancy (8).”
(1) BJOG: An International Journal of Obstetrics and Gynaecology. Danish studies suggest low and moderate drinking in early pregnancy has no adverse effects on children aged 5. June 20, 2012. http://www.bjog.org/view/0/index.html
(2) Kesmodel, UW., Bertrand, J, Stovring, H., Skarpness, B, Denny, CH, Mortensen, EL, and the Lifestyle Pregnancy Study Group (2012). The effect of different alcohol drinking patters in early to mid pregnancy on the child’s intelligence, attention, and executive function. BJOG: An International Journal of Obstetrics & Gynaecology. Doi: 10,1111/j.1471-0528.2012.03393.x
(3) Underbjerg, M., Kesmodel, U., Landro, N., Bakketeig, L., Grove, J., Wimberley, T., Kilburn, T., Svaerke, C., Thorsen, P. and Mortensen, E. (2012). The effects of low to moderate alcohol consumption and binge drinking in early pregnancy on selective and sustained attention in 5-year-old children. BJOG: An International Journal of Obstetrics & Gynaecology. Doi: 10,1111/j.1471-0528.2012.03396x.
(4) Wimberley, T., Grove, J. and Kesmodel, U. (2012). The effects of low t moderate prenatal alcohol exposure in early pregnancy on IQ in 5-year-old children. BJOG: An International Journal of Obstetrics & Gynaecology. Doi: 10.1111/j.1471-0528.2012.03394.x
(5) Kesmodel, US., Underbjerg, M., Kilburn, TR., Bakketeig, L., Mortensen, EL., Landro, NI., Schendel, D., Bertrand, J., Grove, J., Ebrahim, S., & Thorsen, P. (2010). Lifestyle during pregnancy: Neurodevelopmental effects at 5 years of age. The design and implementation of a prospective follow-up study. Scandinavian Journal of Public Health, 2010,; 38: 208-219.
(7) NOFAS, Response to Media Coverage of New Research . www.nofas.org. Retrieved July 7, 2012.
(8) Andersen, A-M N., Andersen, PK., Olsen, J., Gronbaek, M. and Strandberg-Larsen, K. (2011). Moderate alcohol intake during pregnancy and risk of fetal death. International Journal of Epidemiology. 41 (2): 405-413. Doi: 10.1093/ije/dyr189. Ije.oxfordjournals.org
(9) Higgins, JPT., White, IR., and Wood, A.M. (2008). Imputation methods for missing outcome data in meta-analysis of clinical trials. Clinical Trials, 5(3); 225-239. Doi: 10.1177/1740774508091800.