Our thanks to our colleagues at the Telethon Institute for Child Health Research who supported this research. This research was funded by the Health Promotion Foundation of Western Australia (Healthway#10563), NHMRC Program Grants 003209 and 353514. JK is partially funded by a grant to the NPEU and by a National Public Health Career Scientist Award from the Department of Health and NHS R&D (PHSC 022) and CB by NHMRC Fellowships 172303 and 353628.
Reprint requests: Lyn Colvin, MPH, Telethon Institute for Child Health Research, PO Box 855, West Perth WA 6872, Perth, Australia; Fax:61 8 9489 7700; E-mail: firstname.lastname@example.org
Background: High alcohol intake in pregnancy has been linked to abnormal fetal development. There are limited published data in Australia on standard drinks of alcohol consumed on a typical occasion during the periconceptional period or pregnancy.
Methods: During 1995 to 1997, a 10% random sample of all nonindigenous women giving birth in Western Australia was surveyed 12 weeks after delivery (N=4,839). Women were asked questions about alcohol consumption in each of the 4 time periods: the 3 months before pregnancy and each trimester of pregnancy. Questions were framed to measure volume, frequency, and type of alcoholic beverage.
Results: 46.7% of the women had not planned their pregnancy. Most women (79.8%) reported drinking alcohol in the 3 months before pregnancy, with 58.7% drinking alcohol in at least 1 trimester of pregnancy. The proportion of women consuming 1 to 2 drinks on a typical occasion did not change much during pregnancy, but the number of occasions declined. Although the proportion of women consuming more than 2 standard drinks on a typical occasion declined after the first trimester, 19.0% of women consumed this amount in at least 1 trimester of pregnancy and 4.3% of women consumed 5 or more standard drinks on a typical occasion in at least 1 trimester of pregnancy. In the first trimester of pregnancy, 14.8% of women drank outside the current Australian guideline for alcohol consumption in pregnancy, decreasing to 10% in the second and third trimesters.
Conclusions: Women generally reduced their average alcohol consumption and the number of standard drinks on a typical occasion as their pregnancy progressed, although 10 to 14% were drinking outside current guidelines for pregnancy. It is important that all women of child-bearing age are aware, well before they consider pregnancy, of the risks of drinking alcohol during pregnancy so they can make informed decisions about their alcohol consumption in pregnancy.
The threshold for “risky” drinking during pregnancy remains a subject of controversy with some observers finding adverse effects with average levels of exposure as low as 1 drink per week, while others have not seen effects at such low levels (Casswell et al., 2002; Jacobson and Jacobson, 1994; Sood et al., 2001). This controversy is reflected in different guidelines for alcohol consumption during pregnancy in different countries. In 1999, Austria, Canada, Denmark, Ireland, Sweden, and the United States recommended abstinence, while New Zealand and the United Kingdom allowed occasional drinking of low to moderate amounts of alcohol (International Center for Alcohol Policies, 1999). In Australia, until 2001, the national guidelines stated that “responsible drinking during pregnancy must still be considered abstinence” (National Health and Medical Research Council, 1992).
There is little contemporary information in Australia on how many women drink alcohol during pregnancy, how much alcohol is consumed in each trimester, and the type of alcoholic beverages they drink. A study conducted in the 1980s collected information from women during the antenatal period on the type and amount of alcoholic drinks consumed as well as when they were consumed (Walpole et al., 1990). The study found that 49% were abstainers or light drinkers (<0.1 standard drinks/d), 38% were moderate drinkers (0.1–<2 standard drinks/d), and 13% were considered to be heavy drinkers (2 or more standard drinks/d). Among 419 women currently pregnant in a recent cross-sectional study of Australian women aged 18 to 23 years, 24% were nondrinkers, 49% rarely drank, 24% drank at low-risk levels (up to 2 standard drinks/d), and 3% at risky or high-risk levels (more than 2 standard drinks/d; Young et al., 2005). However, this study was not able to determine consumption in terms of standard drinks or the type of alcoholic beverage consumed and did not include pregnant women over the age of 23 years.
Timing of alcohol consumption in relation to awareness of pregnancy is important. Such knowledge is fundamental to the design of targeted interventions for the prevention of FAS and FASD (Floyd et al., 1999).
The objective of this paper is to report on consumption of alcohol during pregnancy in frequency, timing, type, and amount of alcohol, in a random sample of nonindigenous women in Western Australia who had recently given birth when the data were collected. Prepregnancy consumption of alcohol is also reported, including in relation to whether the pregnancy was planned. Finally, results are viewed with reference to the current national guideline for alcohol consumption during pregnancy, as the threshold for risky alcohol consumption in pregnancy is still a matter for debate internationally.
MATERIALS AND METHODS
During 1995 to 1997, a 10% random sample selected from the Western Australia Midwives' Notification System of all women giving birth in Western Australia was invited to participate in the “West Australian Pregnancy and Infancy Survey.” Details of the survey have previously been described (Kurinczuk et al., 1999). The infants were born in the period January 1, 1995 to June 30, 1997. Mothers who delivered a stillborn infant (N=20), whose baby died during the first 12 weeks of life (N=20), and those relinquishing their baby for adoption (estimated at <5) were not eligible for inclusion. Metropolitan indigenous mothers (N=180) were participating in a similar but more culturally appropriate study running concurrently, and were not invited to participate in the Pregnancy and Infancy Survey. Four thousand eight hundred sixty-one women (80.8%) completed a self-administered postal questionnaire 12 weeks postpartum that contained questions on health-related behaviors before, during, and after pregnancy. Nonresponders to the initial mailing were sent a second full mailing and those failing to respond to this were contacted by telephone. When the data from the responders were compared with data available about all births in Western Australia (Stanley et al., 1997), those women who responded were representative of all nonindigenous women with live births in that period, with the exception of a slight under-representation of mothers with low–birth-weight infants (<2,500g; 5.3% overall vs 4.2% in the sample) and mothers who were <20 years old (6% overall vs 3.8% in the sample).
In addition to other questions, women were asked whether at any stage in their life they had ever drunk alcohol and, if yes, to indicate in separate tables for each of 4 time periods (for the 3-month period before the pregnancy and for each trimester of the pregnancy) how often they drank alcohol (5 or more days per week; 3–4; 1–2 d/wk; 1–2 d/mo; less than once a month; and never), and the quantity consumed (e.g., number of cans, glasses, bottles) on a typical occasion for each of 4 types of alcoholic beverage (beer; wine/champagne; spirits/liqueurs; and fortified wines; Fig. 1). Of the 4,861 questionnaires received, 22 (0.4%) did not respond to the questions relating to alcohol consumption, resulting in a sample size of 4,839 women.
The 4 time periods were described in the questionnaire as:
(i)“During the 3 months before your recent pregnancy …”;
(ii)“During the first 3 months (1–3 months or 1–13 weeks) of your recent pregnancy …”;
(iii)“During the middle 3 months (4–6 months or 14–26 weeks) of your recent pregnancy …”; and,
(iv)“During the final 3 months (7–9 months or 27–40 weeks) of your recent pregnancy …”
We calculated 2 measures of alcohol consumption for each time period:
(i)Average daily and weekly consumption: The survey questions required a response to each type of alcoholic beverage within each 3-month period, so average daily consumption was calculated as the sum of the number of standard drinks of each type of alcoholic beverage during the period, divided by 91 days, and weekly consumption was calculated as 7 times the daily consumption.
(ii)Consumption on a “typical occasion”: The volume of alcohol consumed in standard drinks on a typical occasion was not aggregated across types of alcoholic beverages as the questions were not framed in a way that differentiated between types of alcoholic beverage and typical occasions. The maximum number of standard drinks consumed on a typical occasion for any one of the types of alcoholic beverage was used instead. For those women who drank more than 1 type of alcoholic beverage during the time period, these results are a minimum estimate.
Standard drink calculations were based upon Australian measures (Fig. 2) and derived during the data analysis stage, not by the study respondents. The survey questions did not differentiate between the strengths of beer so the standard drink size was calculated assuming full-strength beer had been consumed as this was the most common type of beer consumed by women in that period (Bower et al., 2004).
A conservative approach was used in interpretations of the data for the 2 consumption measures when responses were unclear. For example, where respondents used a tick in a column as a response instead of indicating numeric values for alcohol consumption frequencies, amounts, or drink size, the response was set at the minimum level of 1 measure for that type of alcoholic beverage and time period. If the alcohol consumption frequency was not indicated then the minimum frequency of 0.5 d/mo was applied. Consumption frequency calculations used the lower of the days marked from the response: for example, for a response marked “3 to 4 d/wk” then 3 d/wk was used as the frequency value. If a range was given for the “number of drinks on a typical occasion,” it was taken to be the higher of the range as participants almost always under-report their consumption (Del Boca and Darkes, 2003; Stockwell et al., 2004). A small number (about 15%) provided responses as a range. For example, a response of “1 to 2 cans of beer, 3 to 4 d/wk” was coded as 2 cans, 3 d/wk then converted to 3 standard drinks/d, 9 standard drinks/wk. (1 can of full-strength beer is equivalent to 1.5 standard drinks in Australia; Fig. 2.)
Data analyses were restricted to those records where valid responses to the alcohol consumption questions were available (N=4,839). Data were analyzed for the 4 time periods: the 3-month period before pregnancy and each of the 3 trimesters of pregnancy. If the pregnancy resulted in a preterm birth, then the woman was not included in the last trimester analyses (N=31, mean gestational age <32 weeks). SAS/STAT version 8.2 software was used for the statistical analyses (SAS/STAT, 1999–2001).
A new alcohol guideline was introduced into Australia in 2001 after this study was conducted (Fig. 3). Analyses were also performed to categorize women by compliance with this guideline, had it been in operation at the time of the survey. Alcohol consumption was categorized as “abstain,”“within,” and “outside.” The guideline specifies a minimum of 2 hours as the time frame when drinking more than 2 standard drinks. As the questions in the survey did not specify any time frame and to use the current guideline, we assumed a “typical occasion” to be at least 2 hours.
The mean age at delivery of the 4,839 women was 29.5 years (SD=5.2 years), with a range from 15.0 to 46.8 years. It was their first pregnancy for 29% of the women surveyed. Thirty percent of the women smoked during their pregnancy. The mean and median gestational age was 39 weeks (N=4,918; SD=1.8 weeks; range 24–44 weeks). The mean birth weight was 3,391 g (N=4,920; SD=539 g; range 670–5340 g). Multiple pregnancies accounted for 1.4% of the deliveries. This study is a good random sample of the nonindigenous women giving birth in Western Australia in 1996 when the mean maternal age at delivery was 28.4 years (SD=5.5 years), with a range from 13 to 48 years and the mean birth weight of the infants was 3,340 g (SD=586 g; Gee, 1998). 46.7% of the women had not planned their pregnancy (2,250/4,816; 23 missing responses (0.5%) to this question). 15.9% (N=768) had been trying to become pregnant for <1 month and 37.3% (N=1,798) for more than 1 month. The 1997 cohort of women were also asked when they had first became aware that they were pregnant and of these women, 83.8% were aware they were pregnant by 6 weeks gestation, and 96.7% by 11 weeks.
Alcohol Consumption Within Study Time Periods
8.1% (390/4,839) of the women indicated that they had never consumed alcohol at any stage in their life and a further 10.2% (494/4,839) of the women did not consume alcohol during the entire period of 3 months before and during pregnancy (total abstainers, N=884; 18.3%). The remaining 81.7% (3,955/4,839) of the women drank alcohol at some stage in the 3 months before and/or during pregnancy. 58.7% (2,840/4,839; Table 1) of all women drank alcohol while pregnant; 28.2% (1,363/4,839) drank alcohol during the 3 months before and during each trimester of pregnancy. Interestingly, 1.4% abstained from alcohol in the 3 months before pregnancy and the first trimester but then started drinking in the second or third trimester (data not shown). Only 19.8% (356/1,798) of the women who had been trying to become pregnant for more than 1 month abstained from drinking alcohol in the 3 months before becoming pregnant. This is not significantly different from the 21.3% of women who had not planned their pregnancy and abstained from alcohol during this same period. The difference in alcohol consumption between the women who had planned their pregnancy and those who had not differed between time periods. There was a significant difference in the first trimester where 60.4% (planned) versus 55.0% (unplanned) did not consume alcohol (p-value=0.0002, OR=0.80, 95% CI: 0.71–0.90); an insignificant difference in the second trimester: 56.7% versus 58.5% (p-value =0.2211, OR=1.07, 95% CI: 0.96–1.20), and a significant difference in the third trimester but in the opposite direction to the first trimester: 53.3% versus 56.9% (p-value=0.0128, OR=1.16, 95% CI: 1.03–1.30).
Table 1. Average Daily Consumption of All Alcoholic Beverages by 4,839 Women at Any Time During the Study Period of the 3 Mo Before and During Pregnancy
The volume of alcohol consumed by the study population of 4,839 women in the 3 months before and during their pregnancy as average daily standard drinks is shown in Table 1. Overall, there was a decrease in the average daily consumption. The majority of women (85.8%) averaged <1 standard drink/d in the 3 months before or during pregnancy. There was an increase in the proportion of women abstaining from alcohol once they became pregnant. In the 3-month period before pregnancy, 20.2% of women abstained compared with over 50% in each of the trimesters of pregnancy, although only 41.3% abstained in all 3 trimesters. There were corresponding reductions in the proportions of women consuming alcohol at each of the levels of average daily consumption from before to during pregnancy. About 2% of women drank an average of more than 2 standard drinks/d in at least 1 trimester of pregnancy.
Consumption on a “Typical Occasion”
When the maximum number of standard drinks on a typical occasion for any 1 type of alcoholic beverage was considered (Table 2), we found that 46.5% (2,251/4,839) consumed more than 2 standard drinks on a typical occasion before pregnancy and 19.1% (922/4,839) drank at this level during pregnancy. For 14.2% (689/4,839) of all the women, their consumption on a typical occasion was 5 or more standard drinks of any 1 type of alcohol in the 3 months before pregnancy and 4.3% (209/4,839) drank at this level at some stage while they were pregnant.
Table 2. Minimal Estimatesa of Standard Drinks on a Typical Occasion Consumed by 4,839 Women at Any Time During the Study Period of 3 Mo Before and During Pregnancy Based on the Maximum Amount for Any 1 Type of Alcoholic Beverage
Minimal estimate: as survey questions did not address combinations of alcoholic beverages consumed on a typical occasion, volume of alcohol consumed cannot be summed across beverages for a typical occasion.
Pregnancies of short gestation excluded (N=31).
0 standard drink
<1 standard drink
1–2 standard drinks
>2 to <5 standard drinks
We found that the proportion of women consuming 1 to 2 standard drinks on a typical occasion did not change very much over the course of the pregnancy (26.0, 29.8, and 31.7% in each trimester; Table 2). When the drinking patterns of the women who consumed 1 to 2 standard drinks on a typical occasion were analyzed, the frequency of consumption of alcohol declined during pregnancy. For example, of the women who consumed 1 to 2 standard drinks of beer on a typical occasion, 8.6% consumed beer on more than 2 d/wk during the 3 months before pregnancy and this declined to 4.9% in the last trimester of pregnancy. For wine/champagne drinkers, the proportion declined from 12.0 to 3.4%.
Consumption of Different Types of Alcoholic Beverages
Consumption patterns of the different types of alcoholic beverages were investigated (Table 3). Of those women consuming alcohol in the 3 months before pregnancy (N=3,860), the majority (55.6%) drank more than 1 type of alcoholic beverage. Once pregnant, the majority (65.5%; 1,335/2,038) drank only 1 type of alcoholic beverage. A greater proportion of women consumed wine/champagne than any other beverage in each of the trimesters of pregnancy, with around twice as many women drinking wine/champagne as beer or spirits/liqueurs.
Table 3. Consumption Pattern of Alcoholic Beverages Consumed by the 3,955 Women Who Consumed Alcohol at Any Time During the Study Period
3 mo before, n=3,860 (%)
First trimester, n=2,038 (%)
Second trimester, n=2,054 (%)
Third trimester, n=2,177 (%)
One alcoholic beverage only
Two alcoholic beverages
Wine/champagne, fortified wines
Beer, fortified wines
Spirits/liqueurs, forfied winesAny 2 of wine/champagne, spirits/liqueurs,
The frequency of alcohol consumption for each alcoholic beverage was also analyzed (data not shown). The proportion of women who drank beer on 1 day or more per week decreased between the 3 months before and the third trimester (e.g., 6.7% women consumed beer 3 to 4 d/wk in the 3 months before and this declined to 2.6% in the third trimester). Similarly, the proportion of women who drank wine/champagne on 1 day or more per week decreased. However, the proportion in each frequency classification plateaued during the second and third trimesters: for example, 9.3% women consumed wine/champagne 3 to 4 d/wk in the 3 months before pregnancy and this declined to 3.1% in the second trimester and 3.2% in the third trimester.
Women generally reduced their average weekly consumption of standard drinks of different types of alcohol over the 3 trimesters of pregnancy (Table 4). The average weekly consumption of beer decreased from 2.1 standard drinks during the first trimester to 0.9 during the third trimester. For wine/champagne, the average weekly consumption decreased from 1.4 to 1.0 standard drinks; spirits/liqueurs: 2.1 to 1.0 standard drinks; and fortified wines: 0.6 to 0.3 standard drinks. The majority of women (66.8%) drank wine/champagne (2,643/3,955) in the 3 months before pregnancy, 54.9% drank spirits/liqueurs (2,172/3,955), and 38.0% drank beer (1,501/3,955) in the same period.
Table 4. Average Weekly Consumption and Number of Standard Drinks on a Typical Occasion for Each Type of Alcoholic Beverage
When we categorized alcohol consumption according to the 2001 guideline, a little over half of the women abstained from alcohol during pregnancy and 27.4% drank within the guideline in the first trimester and about a third drank within the guideline in the second and third trimesters. During the first trimester, 14.8% of the women were drinking outside the guideline. During the second and third trimesters, this proportion decreased to just over 10% (Table 5).
Table 5. Consumption of Alcohol by 4,839 Women According to the National Health and Medical Research Council Australian Alcohol Guideline 11 at Any Time During the Study Period
Average daily total consumption of standard drinks during the 3-mo time period ≥1; or, usual consumption of standard drinks of a particular type of alcoholic beverage on a typical occasion during the 3-mo time period >2.
In this study of pregnant nonindigenous West Australian women, we report the types of alcoholic beverages consumed, the frequency of alcohol consumption, the trimester of pregnancy, and the volume of each alcoholic beverage consumed on a “typical occasion.” Most women (79.8%) reported drinking alcohol in the 3 months before becoming pregnant, with 58.7% drinking alcohol in at least 1 trimester of pregnancy. The converse of this is that just over 40% of the women reported that they did not drink any alcohol while pregnant. The proportion of women consuming alcohol periconceptionally is higher than reported in the 1988 National Maternal and Infant Health Survey in the United States (Floyd et al., 1999) where 45.2% of women (3,654/9,559) consumed alcohol in the 3 months preceding pregnancy recognition. While 37.3% (1,798/4,816) of the women in our study had planned their pregnancy for more than 1 month, but of these, only 19.8% (356/1,798) did not drink in the 3 months before pregnancy.
Most women who did drink during pregnancy consumed an average of <1 standard drink/d. Averaging alcohol intake can conceal information about the number of occasions of consumption and the amount consumed per occasion—for example, the difference between drinking 1 drink of alcohol every day of the week, versus 7 drinks on 1 occasion (Abel, 1999; Dawson and Room, 2000)—and this was the picture that emerged when we examined our data. We found that the proportion of women consuming 1 to 2 drinks on a typical occasion did not change very much over the course of the pregnancy, but the number of occasions of drinking alcohol declined during pregnancy. This trend was also found in a recent study of hazardous drinkers in pregnancy (Rayburn et al., 2006). Importantly, the proportion of women consuming more than 2 drinks on a typical occasion declined in the second trimester once they realized they were pregnant, although 19.0% of women consumed this amount in at least 1 trimester of pregnancy, and 4.3% of women consumed 5 or more standard drinks on a typical occasion.
It should be noted that due to the way the data had been collected, we were unable to sum alcohol intake over different beverages on a typical occasion but we know that about a third of women consuming alcohol during pregnancy drank more than 1 type of alcoholic beverage (Table 3). The questions did not allow for analysis when more than 1 type of beverage was consumed on a typical occasion—due to our conservative approach to measurement, we did not wish to assume that on a typical occasion, the pregnant woman consumed each type of alcoholic beverage she reported. Instead, we interpreted the responses when multiple beverages were reported as the usual amount consumed when that particular beverage was typically consumed. Thus, these figures are likely to be minimal estimates.
Results from a recent Russian study of 413 pregnant women visiting prenatal care facilities found that 85.0% reported some alcohol consumption during the month around the time of conception or in the most recent month of pregnancy, and 51.9% of these drinking women reported some alcohol use in the most recent month (Chambers et al., 2006). Although that study was from a far smaller sample than our study, the consumption levels were similar to our results where 81.7% of the women consumed alcohol in at least 1 time period, and 42.2% during the first trimester.
Our results are also consistent with the finding from the report, Statistics on Drug Use in Australia 2004, where nearly half of Australian women continue to drink while pregnant or breastfeeding (Australian Institute of Health and Welfare, 2005). The report is from the 2004 National Drug Strategy Household Survey. Almost all of those surveyed in that study said they reduced their drinking on finding they were pregnant, but the report suggests Australian women are more than 3 times as likely as those in the United States to drink while pregnant. Although there is equivocal evidence for the effectiveness of health warning labels (HWLs) on alcoholic beverages (Kaskutas, 1995; International Center for Alcohol Policies, 1997), it has been mandatory in the United States since 1989 (Kaskutas, 1995)—there are no similar warnings on alcoholic beverages in Australia. Attitudes and knowledge of the risk factors relating to alcohol consumption during pregnancy are an important area of research. Our study contains detailed alcohol consumption data including the frequency and volume of consumption of types of beverages within the 4 time periods. It is expected that this information will be used to assist with specific targets in public health promotion policy relating to the drinking patterns of pregnant women.
The strengths of this study are that a random sample of women giving birth was surveyed and there was a high response for a postal survey. Furthermore, the women were, on the basis of their infants' birth characteristics, a representative sample of nonindigenous Western Australian women, with the exception of a slight under-representation of mothers with low–birth-weight infants (<2,500 g; 5.3% overall vs 4.2% in the sample) and mothers who were <20 years old (6.0% overall vs 3.8% in the sample). As the study dealt with a whole range of pregnancy and early infancy exposures and did not specifically focus on alcohol consumption, we have no a priori reason to suppose that women with atypical drinking patterns would have been more or less likely to participate. Data relating to different types of alcoholic beverages were collected. Previous studies have shown that this increases the estimates of volume consumed (Casswell et al., 2002; Dawson, 1998). It also provides more detailed information on the drinking patterns of pregnant women. The Australian Longitudinal Study on Women's Health (ALSWH) reported results on Australian women and alcohol consumption for the period 1996 to 2003 (Young et al., 2005). That study used 3 age-based cohorts and only the younger group of women aged 18 to 23 years were of child-bearing age. However, this study was not able to determine consumption in terms of standard drinks or the type of alcoholic beverage consumed.
While the study relied on retrospective self-reports, the questions concerning alcohol consumption were designed to enable good quantifiable data to be collected from the women in relation to the timing (before pregnancy and by trimester), the types of alcoholic beverages consumed, the volume as related to the actual size can, bottle or glass, and the frequency of consumption (Dawson and Room, 2000; Greenfield, 2000). When quantifying alcohol consumption, an additional consideration is that drinks poured in the home and elsewhere outside of serving establishments are not likely to adhere to standard sizes, making accurate estimates difficult (Kaskutas and Graves, 2001). The women were not asked to estimate their consumption in “standard drink” sizes. In Australia, the term “standard drink” refers to the amount of an alcoholic drink containing 10 mg of ethanol. In Western Australia, drink measures more commonly have names that consumers use (for example, a “middy” of beer). For this reason, the study questions allowed the respondents to fill in their usual drink size name for each type of alcoholic beverage. These were then translated into standard drinks.
There are some limitations to the study. The women were surveyed 3 months after the birth of their infant and there may be recall errors or bias and, hence, some random and/or biased misclassification of exposure. The underestimation of alcohol intake has been shown to increase as the recall period extends past 1 week (Ekholm, 2004) and hence our study may be underestimating consumption. Studies have shown that participants almost always under-report their consumption (Del Boca and Darkes, 2003; Stockwell et al., 2004). Stockwell et al. (2001) found that the level of drinking reported in the 1998 National Drug Strategy Household survey is consistent with an adult per capita consumption of only 46.5% of that estimated by the Australian Bureau of Statistics for that year on the basis of import, export, and production data. However, a recent study of self-reported alcohol consumption by pregnant women found that retrospective reports 6 months after birth were higher than concurrent reports (Alvik et al., 2006). The authors concluded that this indicated under-reporting during pregnancy rather than over-reporting after pregnancy. Thus, although there may be some under-reporting of alcohol consumption in pregnancy, in our study this may not be as great as expected as the responses were provided retrospectively.
Omissions from this study were specific questions on drinking to the point of intoxication and the time over which alcohol was consumed on a typical occasion. This is not a new issue in studies of alcohol consumption as time frames are not usually examined—questions are often framed as “how many times during the past month did you have 5 or more drinks on an occasion?” (Ebrahim et al., 1999) or the number of drinks “per drinking day” (Rayburn et al., 2006). However, the current Australian guideline stipulates 2 hours as the minimum time frame and this was not measured in our study.
4.3% of women in this study consumed 5 or more standard drinks of 1 alcoholic beverage on a typical occasion during pregnancy. This is higher than a U.S. study that reported that binge drinking among pregnant women in 1995 was 2.9% (Ebrahim et al., 1999). Although drinking in young women in Australia does not appear to have increased, it is already at high-risk levels (Chikritzhs et al., 2003) and there is evidence that such occasions are related to the development of FASD in offspring (O'Leary, 2002). A recent study in New South Wales of hospital admissions during pregnancy found 0.08% of births had admissions for alcohol-related diagnoses (Burns et al., 2006). This study would have only detected the women with a major alcohol problem requiring hospitalization, whereas our study indicates a higher proportion of women are drinking at risky levels during pregnancy.
When we examined the data in relation to the current guideline (National Health and Medical Research Council, 2001), we found that between 10.6 and 14.8% of pregnant women drank outside the guideline depending on which trimester of pregnancy was examined. This result is similar to those of Walpole et al. (1990) where 13% were considered to be heavy drinkers i.e., more than 2 standard drinks/d. These estimates are likely to be conservative because of the limitations discussed. Almost half the women (47.8%) drank outside the guideline in the 3-month period immediately before pregnancy. However, it may be that many of these women may have reduced their consumption during this period in anticipation of pregnancy.
In summary, we have found that the majority of women (58.7%) drank alcohol in pregnancy despite the advice at the time of the study (1995–1997) being abstinence. However, it is encouraging that many women who drank alcohol reduced their alcohol consumption in the first trimester of pregnancy. Hence, with appropriate information, they and others may be able to (further) reduce or abstain from consuming alcohol when they are pregnant or might soon become pregnant. This is especially relevant for the 4.3% of women who drank 5 or more standard drinks on a typical occasion during pregnancy. The challenge is to develop effective health promotion messages to reach women of child-bearing age before they consider pregnancy so they can make informed decisions about reducing their alcohol consumption or abstaining.
Our thanks are due to our colleagues at the Telethon Institute for Child Health Research who provided information. We also thank and gratefully acknowledge the assistance given by Kaye Moore, Tracy Barker, Colleen O'Leary, Meryl Biggs, Jackie Goldfinch, and Katherine Whipp, Telethon Institute for Child Health Research, Perth, Western Australia.