A comparison of the prevalence of prenatal alcohol exposure obtained via maternal self-reports versus meconium testing: a systematic literature review and meta-analysis

Prenatal alcohol exposure (PAE) may cause a number of health complications for the mother and the developing fetus, including Fetal Alcohol Spectrum Disorder (FASD). As a “spectrum” disorder, FASD encompasses a broad array of physical defects, cognitive, behavioural, emotional, and adaptive functioning deficits, as well as congenital anomalies, such as malformations and dysplasia of the cardiac, skeletal, renal, ocular, auditory, and other systems [1]. These impairments are likely to have lifelong implications, which result in a significant economic burden for any society. However, the burden of FASD is not measurable by cost alone, the lifelong hardships faced by these children and their families are also of considerable importance.

Identification of infants exposed to alcohol in utero is crucial, as it can lead to close monitoring of his/her development, facilitate early FASD diagnosis, and implement timely interventions, if necessary. Early interventions have long-term benefits for a child with FASD and can potentially reduce the occurrence of secondary disabilities including poor school performance, addictions, mental health problems, sexually deviant behaviour, dependent living, legal issues, and incarceration [2]. Early FASD diagnosis and providing a stable and nurturing environment for the child have been shown to improve outcome and decrease the risk of secondary disabilities by up to fourfold [2, 3].

Furthermore, screening of neonates for PAE and early FASD diagnosis can prevent subsequent alcohol-exposed births by providing appropriate interventions, treatment, counselling, and support for birth mothers with unrecognized alcohol dependence and mental health problems [1, 4]. Appropriate screening strategies may also facilitate early recognition and intervention for affected siblings.

In order for an infant to be diagnosed with an FASD, PAE needs to be confirmed (with the exception of Fetal Alcohol Syndrome (FAS), which can be diagnosed without PAE confirmation). This can be problematic as maternal self-report data are often under-reported for a variety of reasons (e.g., social desirability bias, recall bias, and/or fear that the child may be taken away). Thus, the detection of PAE in neonates by maternal self-report was shown to be unreliable [5]; thereby necessitating the use of an unbiased biomarker to identify those at risk for FASD.

The use of biological markers has emerged as a practical method for the identification of PAE [6, 7]. Currently, there are several measurable biomarkers available for detecting PAE (i.e., fatty acid ethyl esters [FAEE], ethylglucuronide [EtG], ethlysulphate [Ets], and phosphatidylethanol [PEth]) in a range of neonatal matrices (e.g., hair, meconium, blood, placenta, and umbilical cord) [6, 8‐10]. FAEE in meconium and hair is currently the most commonly used tool to estimate the prevalence of PAE [8]. The validity of the other biomarkers, listed above, remains to be established in neonatal matrices [8].

The detection of FAEE, products of non-oxidative ethanol metabolism, above laboratory cut-points in meconium, has been repeatedly established as a novel biomarker of fetal ethanol exposure in the second and third trimesters of pregnancy [11‐19]. Meconium testing for FAEE is a validated method for detecting PAE; it has been shown to have high sensitivity (84.2%) and specificity (83.3%) [11]. Meconium comprises the neonate’s first several bowel movements, identified most commonly by its dark green/black colour and lack of odour. It is generally agreed upon that meconium formation begins at approximately 12 weeks of gestation (i.e., at the end of the first trimester), when fetal swallowing of amniotic fluid is initiated [20, 21]; however, some researchers have suggested formation begins even later (up to 20 weeks of gestation) [22‐26]. FAEE do not cross the human placenta, causing meconium to serve as a reservoir of fetal chemical exposures during the second and third trimesters of pregnancy. Thus, the presence of FAEE in the meconium is a true reflection of fetal ethanol metabolism [14]. Cumulative meconium FAEE concentrations exceeding 2.0 n▪mol/g, the internationally accepted cut-point [13], may be indicative of seven or more drinks per week or five or more drinks per occasion (i.e., “binge” drinking) [15, 27].

The purpose of the current study was to compare the prevalence of PAE, as reported via maternal self-reports, with the results of meconium testing (the most commonly used laboratory screening method for estimating the prevalence of PAE, with proven accuracy), and to quantify the disparity between these two methods.

The systematic literature review and meta-analyses were conducted and reported according to the standards set out in Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA; http://​www.​prisma-statement.​org/​) [28].

The prevalence estimates of PAE will be, on average, four times higher when using meconium testing compared to the prevalence estimates obtained using maternal self-reports; however, it is not currently feasible to use meconium testing to screen every newborn or to implement a universal meconium testing screening program. There are also legal and ethical issues surrounding meconium testing and its use as a universal screening tool (see Dickens [49]), which need to be resolved before implementation of such a screening program takes place. Thus, a maternal self-report method is the most widely used at the present time.

The current study has limitations due to the lack of consistency across studies with respect to maternal self-report data. For instance, the studies used different instruments/methods to obtain the self-reported data, each with its own sensitivity and specificity. Furthermore, only one study [45] had provided a breakdown by frequency of alcohol use: “daily”, “weekly”, and “monthly” (however, the amount consumed was not reported). In addition, the majority of studies did not specify the time period captured by the exposure assessment, and as a result, it cannot be said definitively that the maternal self-reported data is reflective of the same time period as the meconium testing data (which detects alcohol use in the second and third trimesters only).

Mecounium testing also has several limitations. Due to the inability of meconium testing to ascertain PAE during the first trimester and to “detect” if pregnant women consumed less than seven drinks per week and/or did not practice “binge” drinking (five or more drinks per occasion) during the second and third trimesters of pregnancy, the prevalence of PAE obtained using this method is likely to be an underestimate and, thus, the pooled difference reported in this paper is likely to be underestimated. However, a positive meconium test (i.e., above the cumulative FAEE cut-point) indicated heavy drinking in the second and third trimesters. In countries where the majority of women are aware of the adverse effects of alcohol consumption on the developing fetus (e.g., Australia and Canada [50, 51]), a positive meconium test likely indicates continuous drinking during pregnancy and thus, alcohol dependence; therefore, meconium testing can be said to identify the most at-risk group of neonates.

Additionally, there are variations in the laboratories’ methods for testing meconium for the presence of FAEE (e.g., in the extraction methods used, and in the analytical strategies employed), which affect the PAE prevalence estimates obtained [52]. Furthermore, the association between level of alcohol consumption and the level of FAEE is difficult to ascertain due to individual differences of both the mother and the fetus [52].

Given the limitations of meconium testing, one should err on the side of caution when screening for PAE, as there are potentially serious consequences of both false positive and false negative results. This is especially important when a “positive” test is used for any other purpose than to identify individuals at-risk and in need of intervention, such as when PAE is used as an alternative explanation for poor birth outcomes in medical malpractice claims. A “positive” meconium test result could also lead to increased contact with child protective services in some countries. Accordingly, there is a need for research that can lead to improvements in the sensitivity and specificity of meconium testing so that a universal standard can be definitively established.

The results of this study also reveal that existing estimates of the prevalence of PAE (regardless of the method of detection) vary across countries, as well as within countries. Furthermore, large variations in the estimates of PAE as measured by maternal self-reports were observed when compared to estimates of PAE obtained from meconium testing. Between country variations in the prevalence of PAE likely stem from differences in maternal drinking behaviours, as well as from political, ideological, cultural, and legal differences between countries. As discussed in a recent study by Drabble and colleagues [53], countries vary considerably in the quality and type of public messaging in terms of levels of political attention, responsiveness of governments, and the information relayed to the public regarding alcohol and pregnancy. For countries in which there is heightened political attention, responsiveness, and public messaging regarding alcohol use during pregnancy, there is an increased understanding of the risk associated with alcohol consumption and pregnancy, and consequently, a lower observed prevalence of PAE.

Differences in ideological and cultural conditions between countries can also influence both the prevalence of alcohol use during pregnancy and the accuracy of maternal self-reports [53]. For example, the social acceptability of alcohol use among women influences the stigmatization of women who do consume alcohol, especially during pregnancy [53]. Thus, in those countries where alcohol consumption by women is not socially acceptable, women are less likely to report their use of alcohol during pregnancy. This could explain the results reported for Uruguay [37‐39], where maternal self-reports were not that different from the results of the meconium testing, and could explain the results from some study sites in Italy [45], where maternal self-reports yielded higher prevalence estimates compared to meconium testing. It is possible that in Uruguay and in Italy alcohol consumption by women is less stigmatized when compared to other countries, and thus, women are more likely to accurately report their alcohol consumptions during pregnancy (Magri R 2013, personal communication, May 11). The results from Uruguay and Italy may be explained further by the potential that “low/moderate” levels of alcohol consumption were not “detected” by meconium testing, whereas these levels can be measured by maternal self-reports. Furthermore, a lack of public awareness may exist in Uruguay and in Italy of the deleterious effects of alcohol on a developing fetus, and as a consequence, women in these countries may be more willing to reveal alcohol use during pregnancy [53].

Finally, laws concerning alcohol use during pregnancy may impact variations in self-reported PAE. In Canada, for example, women are not held legally responsible for prenatal injuries, as holding a woman legally responsible in these circumstances is considered to a violation of her fundamental rights [49]. Therefore, an absence of legal consequences of PAE may cause women to be more willing to disclose their alcohol use during pregnancy.

Based on the results of this study, health care professionals should be aware of the under-reporting of PAE by maternal self-reports. If maternal self-reports are not adjusted for under-reporting, a number of infants prenatally exposed to alcohol will not being recognized as such. Although meconium testing provides more accurate data, this paper advocates for the use of meconium testing in addition to maternal self-reports when testing for PAE. However, given that meconium testing cannot determine exposure in the first trimester, further research is needed in order to discover new and validate existing biomarkers for the detection of PAE. As indicated above, there are a number of biomarkers available for detecting PAE in a range of neonatal matrices. These biomarkers need to be further explored and the accuracy (i.e., sensitivity and specificity) of the methods employed needs to be established; however, some preliminary results appear promising (see for example, Bakhireva et al. [54], Matlow et al. [55], and Shukla et al. [56]).