Abstract

The authors tested the hypothesis that the birth prevalence of gastroschisis is positively associated with use of recreational drugs in early pregnancy. A matched case-control study was carried out in three regions of the United Kingdom over the period January 2001 through August 2003. For each case, three liveborn controls were matched by initial intended place of delivery, region, and maternal age. Maternal hair analysis provided independent verification of recreational drug use. Conditional logistic regression was used to estimate mutually adjusted odds ratios. Estimates were revised using data from hair analysis. Statistically significant adjusted odds ratios for gastroschisis were associated with first-trimester use of 1) any recreational drug (odds ratio (OR) = 2.2, 95% confidence interval (CI): 1.2, 4.3) and 2) vasoconstrictive recreational drugs (defined as cocaine, amphetamines, and ecstasy) (OR = 3.3, 95% CI: 1.0, 10.5). Other significant exposures included aspirin use (OR = 20.4, 95% CI: 2.2, 191.5), cigarette smoking (OR = 1.7, 95% CI: 1.1, 2.6), and prior history of gynecologic infection/disease (OR = 2.6, 95% CI: 1.2, 5.6). Recreational drug use is a significant risk factor for gastroschisis and is one of a constellation of potentially preventable exposures which include cigarette smoking, aspirin use, and history of gynecologic infection/disease. Maternal hair analysis proved an acceptable and valuable method of independently verifying recreational drug use.

Over recent decades, epidemiologic studies have reported a steady global increase in the birth prevalence of gastroschisis (1–6), an abdominal wall defect that shows a strong inverse relation with maternal age. Recent data from the British Isles Network of Congenital Anomaly Registers confirm both this rising birth prevalence and the fact that approximately three quarters of all infants with gastroschisis are born to mothers under the age of 25 years (7). The search for the reason(s) for this strong relation with young maternal age has driven research in this area.

Theories concerning the pathogenesis of gastroschisis are varied (8–10), and the defect is thought to develop at 5–9 weeks of gestation. Studies have been carried out to investigate the relation between use of vasoconstrictive drugs, including therapeutic and recreational drugs, and gastroschisis (11–14). The rationale for these investigations was Hoyme et al.'s (10) vascular disruption theory regarding the etiology of gastroschisis, the strong association with young maternal age (15, 16), and the higher prevalence of (and often unplanned and sporadic) use of recreational drugs in the young. These case-control studies have found an increased risk of gastroschisis associated with such drug use but have been limited by their reliance on data from maternal notes and retrospective interviews alone, without any independent verification of recreational drug use. Other reported risk factors include low socioeconomic status, low body mass index, and cigarette and alcohol use in early pregnancy (15, 17–19).

A major issue in the collection of data about risk behaviors during pregnancy is the unwillingness of mothers to admit to such activity. US studies have found high levels of underreporting of recreational drug use during pregnancy in validation studies using hair analysis (20, 21). However, there are no comparable data available from the United Kingdom. In a recent small United Kingdom study of mothers with pregnancies complicated by gastroschisis, hair analysis was used to determine the presence of recreational drugs in maternal hair and the timing of drug use (22). However, this information was not compared with data collected from the mothers.

Here we report results from a large, population-based European case-control study of gastroschisis, with biologic verification of maternal drug use by contemporaneous hair analysis.

MATERIALS AND METHODS

A matched case-control study was carried out to test the hypothesis that the risk of gastroschisis is positively associated with the use of recreational drugs in the first trimester of pregnancy. The study population comprised all deliveries and terminations of pregnancy to mothers resident in the geographic areas of three collaborating regional congenital anomalies registers in the United Kingdom (the Trent, Northern, and West Midlands health regions) over the period January 1, 2001, to August 31, 2003. Cases were defined as all women with gastroschisis-affected pregnancies (International Classification of Diseases, Tenth Revision, code Q79.3) residing in the three geographically defined areas. The majority of gastroschisis case mothers were recruited into the study by their hospital consultant at around 34 weeks' gestation after the mothers had had time to come to terms with the results of their prenatal diagnosis. The remaining cases were recruited by an appropriate health professional following delivery. To form the control group, three mothers of liveborn infants not affected by gastroschisis were matched to each case mother by maternal age (to within 1 year), initial intended place of delivery, and region of maternal residence. Control mothers were recruited into the study by a network of midwives and health visitors across the three regions. Where control mothers declined to join the study, further appropriate women were approached until a full complement of three controls per case was achieved. Written consent was gained from all consenting case and control mothers.

Data were collected for all study subjects by a trained interviewer in each region. The interviewers used a two-part structured questionnaire to collect information from antenatal notes and to conduct a maternal interview that included gathering data on all potential risk and confounding variables. To ensure understanding and standardization of data collection for risk behaviors and use of medications, cards listing categories of therapeutic drugs, types of alcohol, and all known descriptions of recreational drugs were used as prompts. The timing of all reported drug and alcohol use was recorded. If permission for hair sampling was granted, approximately 100 head hairs were collected from the posterior vertex region of the mother's scalp. The hairs were tied together at the scalp end, sealed in a plastic bag, and marked with an identifier and appropriate dates. All consenting case and control mothers were interviewed and had hair samples collected within 6 weeks of delivery; 99 percent of cases interviewed allowed hair to be collected, as did 98 percent of controls.

The rate of hair growth during pregnancy is approximately 0.9 cm per month (23). Using this information, the segment of each hair sample that was growing during the first trimester of pregnancy was prepared for analysis using gas chromatography-mass spectrometry. The laboratory was blinded to both the case/control status of the sample and the participant's reported use of recreational drugs. Samples were prepared by washing and grinding of the hair. This was followed by extraction, elution, and derivatization of the drugs and their metabolites (24). Hair-washing solutions were kept as a double-check for contamination where a positive result was found. The initial gas chromatography-mass spectrometry analysis failed to validate any of the reported drug use, and therefore this analysis was felt to be invalid. Further analyses using radioimmunoassay (21, 25) screening with confirmation by tandem liquid chromatography-mass spectrometry (26, 27) were carried out by the senior laboratory manager. However, because of both financial and time constraints, this additional analysis was only undertaken for a subsample of the women (known recreational drug users from maternal report, for validation purposes, and a random sample of the remaining hair samples) for the detection of vasoconstrictive recreational drugs, defined as cocaine, amphetamines, and ecstasy (3,4-methylenedioxy-N-methylamphetamine) (14). A hair sample was considered to be positive for recreational drugs using the threshold values of Kintz (28) for three ions associated with each drug.

Statistical analysis

Potential risk and confounding factors for gastroschisis were identified using univariate conditional logistic regression. A stepwise conditional logistic regression model, using case-control status as the binary response variable, was constructed in STATA (29), including the exposure variables of primary interest (any recreational drug use and any vasoconstrictive recreational drug use) and all potential risk and confounding variables as covariates. Variables that did not have a statistically significant effect (p < 0.05) on the fit of the data were removed. Models were checked for interactions, the effect of outliers, and the relevant component of influence and leverage. This process was initially carried out using data from maternal interviews alone.

Given that only a portion of the subjects had results from hair analysis (26.3 percent of control mothers and 39.5 percent of case mothers), an unconditional logistic regression analysis was carried out (adjusting for maternal age) excluding subjects without hair analysis results.

The Trent Multi-Centre Research Ethics Committee provided ethical approval for this study. Written consent was given by all participating mothers.

RESULTS

Over the study period, 165 infants with gastroschisis were delivered to 164 eligible mothers—an overall fetal/birth prevalence of 4.24 per 10,000 total births. The majority of these cases were isolated anomalies (n = 154; 93.3 percent). In total, 144 of 164 mothers agreed to participate in the study (a response rate of 87.8 percent). Of these, 142 mothers (98.6 percent) provided a hair sample. There were no statistically significant differences between responding and nonresponding cases in terms of maternal age, ethnic group, plurality, or infant sex. Outcomes among the nonresponding cases were significantly poorer in terms of prematurity, other anomalies, and more terminations of pregnancy (although this was rare: only nine in the total sample). A full complement of 432 control mothers was recruited, with a response to the initial approach of 76.9 percent. Of these 432 control mothers, 423 (97.9 percent) provided a hair sample.

Ten gastroschisis case mothers reported vasoconstrictive recreational drug use in the first trimester of pregnancy: seven ecstasy users, five amphetamine users, and two cocaine users (table 1). Six control mothers reported vasoconstrictive recreational drug use: four ecstasy users, two amphetamine users, and one cocaine user. Univariate analysis showed that case mothers were over 2.5 times more likely than control mothers to be unmarried and to have a history of gynecologic infection or disease and nearly twice as likely as control mothers to be unemployed (table 1). These findings were statistically significant. In addition, case mothers were significantly less likely than controls to own a home (odds ratio (OR) = 0.4, 95 percent confidence interval (CI): 0.3, 0.7), to be overweight (OR = 0.3, 95 percent CI: 0.2, 0.6), or to be obese (OR = 0.2, 95 percent CI: 0.1, 0.6).

TABLE 1.

Unadjusted odds ratios for gastroschisis according to maternal sociodemographic characteristics, medical history, and first-trimester exposures, United Kingdom, January 2001–August 2003

Characteristic or risk factor Cases (n = 144) Controls (n = 432) Unadjusted odds ratio 95% confidence interval 
No. No. 
Ethnic group       
    White 131 91.0 395 91.4 1.0  
    Asian 2.1 11 2.6 0.3 0.1, 1.6 
    Black 1.4 16 3.7 0.8 0.2, 3.2 
    Mixed/other 5.5 10 2.3 1.7 0.4, 7.0 
Parity       
    0 (first pregnancy) 99 68.8 277 64.1 1.7 1.0, 2.9 
    1 24 16.7 108 25.0 1.0  
    ≥2 21 14.5 47 10.9 1.9 1.0, 3.8 
Family history of congenital anomaly 15 10.4 36 8.3 1.3 0.7, 2.6 
History of gynecologic infection/disease* 18 12.5 22 5.1 2.8 1.4, 5.5† 
Marital status       
    Married 21 14.5 117 27.1 1.0  
    Single/separated/divorced 123 85.5 315 72.9 2.6 1.5, 5.5† 
Body mass index‡       
    Normal 103 71.5 260 60.2 1.0  
    Underweight 26 18.1 41 9.5 1.7 1.0, 3.0 
    Overweight 10 6.9 78 18.0 0.3 0.2, 0.6§ 
    Obese 3.5 53 12.3 0.2 0.1, 0.6§ 
Overall social class       
    I, II, or IIInM 56 38.9 195 45.1 1.0  
    IIIM, IV, or V 54 37.5 168 38.9 1.1 0.7, 1.7 
    Unemployed 34 23.6 69 16.0 1.8 1.1, 3.0† 
Homeowner 36 25.0 170 39.4 0.4 0.3, 0.7§ 
Use of a car 95 66.0 306 70.8 0.8 0.5, 1.2 
Exposure to x-rays 2.1 0.7 2.3 0.5, 10.1 
Use of aspirin 4.9 0.5 19.1 2.3, 156.5† 
Use of ibuprofen 6.3 20 4.6 1.4 0.6, 3.1 
Use of phenylpropanolamine 2.1 0.9 2.5 0.5, 12.9 
Use of ephedrine 0.7    
Daily caffeine consumption (cups)       
    0 15 10.4 43 10.0 1.0  
    1–9 75 52.1 243 56.2 0.9 0.5, 1.7 
    ≥10 54 37.5 146 33.8 1.1 0.6, 2.1 
Cigarette smoking 97 67.4 214 49.5 2.1 1.4, 3.2† 
Weekly alcohol consumption (units¶)       
    0 45 31.2 173 40.0 1.0  
    <14 60 41.7 175 40.6 1.3 0.9, 2.1 
    ≥14 39 27.1 84 19.4 1.9 1.1, 3.1† 
Binge drinker 18 12.6 48 11.1 1.2 0.6, 2.1 
Any recreational drug use 24 16.7 24 5.5 3.4 1.9, 6.1† 
Any vasoconstrictive recreational drug use# 10 6.9 1.4 5.7 1.9, 16.7† 
Characteristic or risk factor Cases (n = 144) Controls (n = 432) Unadjusted odds ratio 95% confidence interval 
No. No. 
Ethnic group       
    White 131 91.0 395 91.4 1.0  
    Asian 2.1 11 2.6 0.3 0.1, 1.6 
    Black 1.4 16 3.7 0.8 0.2, 3.2 
    Mixed/other 5.5 10 2.3 1.7 0.4, 7.0 
Parity       
    0 (first pregnancy) 99 68.8 277 64.1 1.7 1.0, 2.9 
    1 24 16.7 108 25.0 1.0  
    ≥2 21 14.5 47 10.9 1.9 1.0, 3.8 
Family history of congenital anomaly 15 10.4 36 8.3 1.3 0.7, 2.6 
History of gynecologic infection/disease* 18 12.5 22 5.1 2.8 1.4, 5.5† 
Marital status       
    Married 21 14.5 117 27.1 1.0  
    Single/separated/divorced 123 85.5 315 72.9 2.6 1.5, 5.5† 
Body mass index‡       
    Normal 103 71.5 260 60.2 1.0  
    Underweight 26 18.1 41 9.5 1.7 1.0, 3.0 
    Overweight 10 6.9 78 18.0 0.3 0.2, 0.6§ 
    Obese 3.5 53 12.3 0.2 0.1, 0.6§ 
Overall social class       
    I, II, or IIInM 56 38.9 195 45.1 1.0  
    IIIM, IV, or V 54 37.5 168 38.9 1.1 0.7, 1.7 
    Unemployed 34 23.6 69 16.0 1.8 1.1, 3.0† 
Homeowner 36 25.0 170 39.4 0.4 0.3, 0.7§ 
Use of a car 95 66.0 306 70.8 0.8 0.5, 1.2 
Exposure to x-rays 2.1 0.7 2.3 0.5, 10.1 
Use of aspirin 4.9 0.5 19.1 2.3, 156.5† 
Use of ibuprofen 6.3 20 4.6 1.4 0.6, 3.1 
Use of phenylpropanolamine 2.1 0.9 2.5 0.5, 12.9 
Use of ephedrine 0.7    
Daily caffeine consumption (cups)       
    0 15 10.4 43 10.0 1.0  
    1–9 75 52.1 243 56.2 0.9 0.5, 1.7 
    ≥10 54 37.5 146 33.8 1.1 0.6, 2.1 
Cigarette smoking 97 67.4 214 49.5 2.1 1.4, 3.2† 
Weekly alcohol consumption (units¶)       
    0 45 31.2 173 40.0 1.0  
    <14 60 41.7 175 40.6 1.3 0.9, 2.1 
    ≥14 39 27.1 84 19.4 1.9 1.1, 3.1† 
Binge drinker 18 12.6 48 11.1 1.2 0.6, 2.1 
Any recreational drug use 24 16.7 24 5.5 3.4 1.9, 6.1† 
Any vasoconstrictive recreational drug use# 10 6.9 1.4 5.7 1.9, 16.7† 
*

Defined as recurrent urinary tract infection, Chlamydia infection, and/or abnormal cervical smear.

Significantly elevated risk in gastroschisis mothers.

Weight (kg)/height (m)2.

§

Significantly reduced risk in gastroschisis mothers.

A United Kingdom standard unit of alcohol is 10 ml or 8 g of pure alcohol. This is equivalent to a half pint of beer, one sixth of a gill of spirits, or a glass of wine (125 ml).

#

Defined as cocaine, amphetamines, or ecstasy.

After adjustment for factors that had a significant effect on the conditional logistic regression model (table 2), mothers who reported using any recreational drugs in the first trimester were over twice as likely to have a baby with gastroschisis as mothers who did not (adjusted OR = 2.2, 95 percent CI: 1.2, 4.3). First-trimester reported use of vasoconstrictive recreational drugs was associated with over a threefold risk of gastroschisis (OR = 3.3, 95 percent CI: 1.0, 10.5). Use of aspirin (OR = 20.4, 95 percent CI: 2.2, 191.5) and cigarette smoking (OR = 1.7, 95 percent CI: 1.1, 2.6) in the first trimester were also significantly associated with increased risk of gastroschisis, as was a history of gynecologic infection or disease prior to the current pregnancy (OR = 2.6, 95 percent CI: 1.2, 5.6).

TABLE 2.

Mutually adjusted odds ratio (from conditional logistic regression) for gastroschisis according to reported maternal use of any recreational drug or vasoconstrictive recreational drug during the first trimester of pregnancy, United Kingdom, January 2001–August 2003

Risk factor (yes:no unless otherwise specified) Adjusted* odds ratio 95% confidence interval 
Use of any recreational drug 2.2 1.2, 4.3† 
Use of a vasoconstrictive‡ recreational drug 3.3 1.0, 10.5† 
Body mass index§   
    Underweight:normal 2.0 1.1, 3.7† 
    Overweight:normal 0.3 0.2, 0.7¶ 
    Obese:normal 0.3 0.1, 0.8¶ 
Marital status (single/separated/divorced:married) 1.8 1.0, 3.3 
Use of aspirin 20.4 2.2, 191.5† 
Homeowner 0.5 0.3, 0.9¶ 
History of gynecologic infection/disease 2.6 1.2, 5.6† 
Cigarette smoking 1.7 1.1, 2.6† 
Risk factor (yes:no unless otherwise specified) Adjusted* odds ratio 95% confidence interval 
Use of any recreational drug 2.2 1.2, 4.3† 
Use of a vasoconstrictive‡ recreational drug 3.3 1.0, 10.5† 
Body mass index§   
    Underweight:normal 2.0 1.1, 3.7† 
    Overweight:normal 0.3 0.2, 0.7¶ 
    Obese:normal 0.3 0.1, 0.8¶ 
Marital status (single/separated/divorced:married) 1.8 1.0, 3.3 
Use of aspirin 20.4 2.2, 191.5† 
Homeowner 0.5 0.3, 0.9¶ 
History of gynecologic infection/disease 2.6 1.2, 5.6† 
Cigarette smoking 1.7 1.1, 2.6† 
*

Adjusted for all other characteristics in the table.

Significantly elevated risk in gastroschisis mothers.

Defined as cocaine, amphetamines, or ecstasy.

§

Weight (kg)/height (m)2.

Significantly reduced risk in gastroschisis mothers.

In all instances where a mother had reported vasoconstrictive recreational drug use in early pregnancy and hair analysis was carried out, hair analysis results concurred. Three (2.1 percent) additional case mothers (two cocaine users and one amphetamine user) and six (1.4 percent) control mothers (three cocaine users, one ecstasy user, one amphetamine user, and one user of both cocaine and ecstasy) were identified as vasoconstrictive recreational drug users upon liquid chromatography-mass spectrometry hair analysis. Unconditional logistic regression analysis (excluding subjects without hair analysis results) indicated that mothers who were deemed to be users of vasoconstrictive recreational drugs in early pregnancy were twice as likely to have a baby with gastroschisis as mothers who were deemed nonusers (adjusted OR = 2.0, 95 percent CI: 0.8, 5.0) on the basis of self-reports or hair analysis.

DISCUSSION

To our knowledge, this study is the largest population-based European case-control study to have investigated the etiology of gastroschisis. Results indicated that mothers who used vasoconstrictive recreational drugs during the first trimester of pregnancy were more likely to have a fetus/baby with gastroschisis than mothers who did not. This corroborates the results from other studies using maternal self-report data (14, 18) and provides additional confirmation of this association using hair analysis. However, while the risk of gastroschisis from maternally reported data was statistically significantly increased more than threefold, unconditional logistic regression analysis (excluding subjects without hair analysis results) reduced the risk to twofold and the result was not statistically significant, probably because of the reduced power of the smaller sample. Nevertheless, our results suggest that the true risk of gastroschisis associated with maternal use of vasoconstrictive recreational drugs in early pregnancy lies somewhere between these two values.

Research suggests that mothers who have poor perinatal outcomes may report adverse exposures more accurately than those who have successful pregnancies, leading to differential misclassification of adverse exposures between cases and controls (30, 31) and an overestimate of the odds ratio. This may be motivated by mothers' need to establish for themselves the cause of the poor outcome, whereas the mothers of healthy infants have no such motivation. In order to overcome this problem, some researchers have suggested using controls with other congenital anomalies (30, 31). Others have suggested that concerns about recall bias in birth defect studies are overrated (32) and that interpreting the results of studies using malformed controls is problematic (33). In this study we decided to use healthy controls, since we wanted to produce results that were applicable to the whole population at risk of gastroschisis. Our independent assessment of maternal drug use using hair analysis attempted to address the issue of recall bias concerning recreational drug use. In order to minimize recall bias for other exposure data, we interviewed all mothers within 6 weeks of their expected date of delivery.

Identification of all gastroschisis cases over the study period was established using a wide network of health professionals and the three regional congenital anomalies registers, all of which use multiple source ascertainment. All cases were confirmed by a neonatologist or pediatric surgeon or by autopsy before inclusion. The response to the study was high: 87.8 percent for case mothers and 76.9 percent for control mothers. The gastroschisis cases associated with nonresponding mothers were more likely to have poorer outcomes and/or be associated with other anomalies. Access to some of these mothers was denied by their obstetrician. However, it is unlikely that the decision of a clinician, based on a mother's welfare and leading to nonresponse, would be associated with the mother's use of recreational drugs or other exposures. The slightly lower response rates for control mothers were partly related to the level of motivation of the health professionals involved in arranging access to the mothers for the interviewers. We validated recreational drug use with hair analysis in order to confirm maternal recall and to identify any underreporting, although initial laboratory failure meant that we were unable to analyze all of the samples collected. Interviewers underwent rigorous training to ensure standardization of data collection between cases and controls, including being shadowed by the principal investigator for some interviews. Every effort was made in the study design and execution to ensure robustness.

In order to prevent confounding by maternal age, controls were successfully matched to cases by age (to within 1 year). Putative risk factors for gastroschisis were included within the mutually adjusted analyses. The final models confirmed previous findings that maternal exposure to vascular disruptors in the form of cigarettes and other vasoconstrictive drugs in early pregnancy increases the risk of gastroschisis (14, 16, 17, 34), an anomaly with a suspected etiology of vascular disruption (10). General therapeutic aspirin use in the first trimester was associated with a 20-fold increased risk of gastroschisis (although this finding was based on only seven case mothers and two control mothers), confirming previous work (12). In addition, associations with social deprivation factors (14, 18) and low body mass index (35) were confirmed. A possible new risk factor, history of gynecologic infection or disease, was identified. Mothers were classified as having a history of gynecologic infection or disease from their past medical history, prior to the current pregnancy, and as such prior treatment for these conditions is unlikely to have confounded this finding. Rather, we believe that this elaborates the description of the lifestyle of mothers who are at risk of a gastroschisis-affected pregnancy, rather than having a direct biologic effect. However, despite previous work showing a significantly increased risk of gastroschisis with alcohol consumption and binge drinking (15, 18), this was not confirmed in our mutually adjusted analysis.

A unique aspect of this study was the validation of maternal interview data concerning recreational drug use by means of hair analysis. Hair analysis using an initial radioimmunoassay screen followed by confirmation by either liquid chromatography-mass spectrometry or gas chromatography-mass spectrometry has been shown to have a high sensitivity (26, 27). In this study, hair analysis results were positive for all reported vasoconstrictive recreational drug use in both case mothers and control mothers where hair samples were suitable for testing. Because of a number of constraints, only a subsample of maternal hair samples was analyzed in this study. Multiple imputation methods were used to predict the likelihood of mothers' taking vasoconstrictive recreational drugs in early pregnancy, where no hair analysis result was available and no such drug use was reported in the maternal interview. However, since this technique is usually used to predict small amounts of missing data in multiple variables, it was felt that any estimate would be associated with high levels of uncertainty, as more than two thirds of hair analysis data was unavailable in this study. While multiple imputation did not substantially affect the findings of the study, the resulting confidence intervals were wide and no longer statistically significant. Therefore, an unconditional logistic regression analysis (excluding subjects without hair analysis results) was carried out for additional rigor. This showed an increased risk of gastroschisis associated with vasoconstrictive recreational drug use, although the risk was only twofold as opposed to the threefold risk observed in the complete data set. Because mothers were fully aware of the reason for the collection of a sample of their hair, it seems unlikely that frequent use would be underreported. However, sporadic vasoconstrictive drug use may have been underreported and subsequently picked up by the hair analysis. Previous work by Kuhn et al. (36) showed a reduction in the risk of cocaine use associated with intrauterine growth retardation resulting from the ability of hair analysis to identify light users of cocaine. As such, the addition of light users of cocaine reduced the odds ratio due to the positive dose-response relation between increasing levels of cocaine use and the outcome. Therefore, the reduction in the odds ratio resulting from our unconditional logistic regression analysis could be explained by differential underreporting of sporadic vasoconstrictive drug use between gastroschisis case mothers and control mothers. Thus, these additional vasoconstrictive drug users identified upon hair analysis may not have taken the drug at the critical developmental point in time or have reached the threshold of risk.

This study provides additional confirmation of the continuing increase in the birth prevalence of gastroschisis in the United Kingdom, with a rate of 4.24 per 10,000 total births, which is in excess of other published United Kingdom data (1, 2, 5). Recent work from a Europe-wide register of congenital anomalies suggested that this increase has occurred across all maternal ages (37), and in Hawaii, Forrester et al. (38) concluded from routinely collected registry data that this increase is not associated with increased recreational drug use. Our findings indicate, however, a greater than twofold risk of gastroschisis associated with vasoconstrictive recreational drug use and a similar risk associated with smoking and other lifestyle factors that are particularly common among younger mothers in the United Kingdom and other developed countries. Therefore, the public health message from this study is to target these risk factors for reduction among young women before they become pregnant in order to prevent any further increase in the birth prevalence of gastroschisis.

Abbreviations

    Abbreviations
  • CI

    confidence interval

  • OR

    odds ratio

This project was funded by National Health Service Research and Development (Executive), Trent, West Midlands, and Northern health regions. Dr. Draper was supported by the Leicester City and County primary care trusts. Dr. Rankin was supported by a Personal Award Scheme Career Scientist Award from the National Institute of Health Research, United Kingdom Department of Health. Dr. Kurinczuk was partially supported by a National Public Health Career Scientist Award (PHCS022) from the Department of Health and National Health Service Research and Development. The National Perinatal Epidemiology Unit at the University of Oxford is supported by the Department of Health (England).

The authors thank all of the health professionals who assisted with the conduct of the study. In addition, they thank the Department of Chemical Pathology at the University Hospitals of Leicester NHS Trust for carrying out the hair analysis and the staff of the Trent, West Midlands, and Northern congenital anomalies registers for the provision and validation of data. In particular, the authors thank Caroline Lamming, Clare Shuter, Theresa Regis, and Gill Thompson for interviewing the mothers and collecting the hair samples and Helen Holden for data preparation. The authors also thank Dr. Mike Wyldes for his clinical input and Prof. Paul Burton for his advice concerning model-checking.

The funding organization played no role in the design and conduct of this study, in the collection, management, analysis, and interpretation of the data, or in the preparation, review, and approval of the manuscript. Dr. Draper had full access to all of the data in the study and takes full responsibility for the integrity and accuracy of the data. The views expressed in this paper are those of the authors and not necessarily those of the funding organization.

Conflict of interest: none declared.

References

1.
Baird
PA
MacDonald
EC
An epidemiologic study of congenital malformations of the anterior abdominal wall in more than half a million consecutive live births
Am J Hum Genet
 , 
1981
, vol. 
33
 (pg. 
470
-
8
)
2.
Tan
KH
Kilby
MD
Whittle
MJ
, et al.  . 
Congenital anterior abdominal wall defects in England and Wales 1987 –93: retrospective analysis of OPCS data
BMJ
 , 
1996
, vol. 
313
 (pg. 
903
-
6
)
3.
Roeper
PJ
Harris
J
Lee
G
, et al.  . 
Secular rates and correlates for gastroschisis in California (1968 –1977)
Teratology
 , 
1987
, vol. 
35
 (pg. 
203
-
10
)
4.
Forrester
MB
Merz
RD
Epidemiology of abdominal wall defects, Hawaii, 1986 –1997
Teratology
 , 
1999
, vol. 
60
 (pg. 
117
-
23
)
5.
Rankin
J
Dillon
E
Wright
C
Congenital anterior abdominal wall defects in the north of England, 1986 –1996: occurrence and outcome
Prenat Diagn
 , 
1999
, vol. 
19
 (pg. 
662
-
8
)
6.
European Surveillance of Congenital Anomalies (EUROCAT)
EUROCAT report 8: surveillance of congenital anomalies 1980–1999. Newtonabbey
2002
Northern Ireland
EUROCAT Central Registry, University of Ulster
7.
United Kingdom Department of Health
Gastroschisis: a growing concern. In: On the state of the public health: annual report of the Chief Medical Officer 2004.
2005
London, United Kingdom
The Stationery Office
(pg. 
41
-
7
)
8.
Duhamel
B
Embryology of exomphalos and allied malformations
Arch Dis Child
 , 
1963
, vol. 
38
 (pg. 
142
-
7
)
9.
deVries
PA
The pathogenesis of gastroschisis and omphalocele
J Pediatr Surg
 , 
1980
, vol. 
15
 (pg. 
245
-
51
)
10.
Hoyme
HE
Jones
MC
Jones
KL
Gastroschisis: abdominal wall disruption secondary to early gestational interruption of the omphalomesenteric artery
Semin Perinatol
 , 
1983
, vol. 
7
 (pg. 
294
-
8
)
11.
Torfs
CP
Katz
EA
Bateson
TF
, et al.  . 
Maternal medications and environmental exposures as risk factors for gastroschisis
Teratology
 , 
1996
, vol. 
54
 (pg. 
84
-
92
)
12.
Werler
MM
Sheehan
JE
Mitchell
AA
Maternal medication use and risks of gastroschisis and small intestinal atresia
Am J Epidemiol
 , 
2002
, vol. 
155
 (pg. 
26
-
31
)
13.
Werler
MM
Mitchell
AA
Shapiro
S
First trimester maternal medication use in relation to gastroschisis
Teratology
 , 
1992
, vol. 
45
 (pg. 
361
-
7
)
14.
Werler
MM
Sheehan
JE
Mitchell
AA
Association of vasoconstrictive exposures with risks of gastroschisis and small intestinal atresia
Epidemiology
 , 
2003
, vol. 
14
 (pg. 
349
-
54
)
15.
Werler
MM
Mitchell
AA
Shapiro
S
Demographic, reproductive, medical, and environmental factors in relation to gastroschisis
Teratology
 , 
1992
, vol. 
45
 (pg. 
353
-
60
)
16.
Haddow
JE
Palomaki
GE
Holman
MS
Young maternal age and smoking during pregnancy as risk factors for gastroschisis
Teratology
 , 
1993
, vol. 
47
 (pg. 
225
-
8
)
17.
Goldbaum
G
Daling
J
Milham
S
Risk factors for gastroschisis
Teratology
 , 
1990
, vol. 
42
 (pg. 
397
-
403
)
18.
Torfs
CP
Velie
EM
Oechsli
FW
, et al.  . 
A population-based study of gastroschisis: demographic, pregnancy, and lifestyle risk factors
Teratology
 , 
1994
, vol. 
50
 (pg. 
44
-
53
)
19.
Lam
PK
Torfs
CP
Brand
RJ
A low pregnancy body mass index is a risk factor for an offspring with gastroschisis
Epidemiology
 , 
1999
, vol. 
10
 (pg. 
717
-
21
)
20.
Shiono
PH
Klebanoff
MA
Nugent
RP
, et al.  . 
The impact of cocaine and marijuana use on low birth weight and preterm birth: a multicenter study
Am J Obstet Gynecol
 , 
1995
, vol. 
172
 (pg. 
19
-
27
)
21.
Kline
J
Ng
SK
Schittini
M
, et al.  . 
Cocaine use during pregnancy: sensitive detection by hair assay
Am J Public Health
 , 
1997
, vol. 
97
 (pg. 
352
-
8
)
22.
Morrison
JJ
Chitty
LS
Peeble
D
, et al.  . 
Recreational drugs and fetal gastroschisis: maternal hair analysis in the peri-conceptual period and during pregnancy
Br J Obstet Gynaecol
 , 
2005
, vol. 
112
 (pg. 
1022
-
5
)
23.
Pecoraro
V
Barman
JM
Astore
I
The normal trichogram of pregnant women
1969
Oxford, United Kingdom
Pergamon Press
pg. 
9
 
24.
Kintz
P
Cirimele
V
Interlaboratory comparison of quantitative determination of amphetamine compounds in hair samples
Forensic Sci Int
 , 
1997
, vol. 
84
 (pg. 
151
-
6
)
25.
Spiehler
V
Hair analysis by immunological methods from beginning to 2000
Forensic Sci Int
 , 
2000
, vol. 
107
 (pg. 
249
-
59
)
26.
Moller
MR
Fey
P
Rimbach
S
Identification and quantification of cocaine and its metabolites, benzoylecgonine and ecgonine methyl ester, in hair of Bolivian coca chewers by gas chromatography/mass spectrometry
J Anal Toxicol
 , 
1992
, vol. 
16
 (pg. 
291
-
6
)
27.
Ostrea
EM
Knapp
DK
Tannenbaum
L
, et al.  . 
Estimates of illicit drug use during pregnancy by maternal interview
J Pediatr
 , 
2001
, vol. 
138
 (pg. 
344
-
8
)
28.
Kintz
P
Moffatt
AC
Osselton
MD
Widdop
B
Hair analysis.
Clarke's analysis of drugs and poisons
 , 
2004
3rd ed
London, United Kingdom
Pharmaceutical Press
(pg. 
124
-
33
)
29.
Stata Corporation.
Stata statistical software: release 7.0
2001
College Station, TX
Stata Corporation
30.
Werler
MM
Pober
BR
Nelson
K
, et al.  . 
Reporting accuracy among mothers of malformed and nonmalformed infants
Am J Epidemiol
 , 
1989
, vol. 
129
 (pg. 
415
-
21
)
31.
Lieff
S
Olshan
AF
Werler
M
, et al.  . 
Selection bias and the use of controls with malformations in case-control studies of birth defects
Epidemiology
 , 
1999
, vol. 
10
 (pg. 
238
-
41
)
32.
Khoury
MJ
James
JM
Erickson
JD
On the use of affected controls to address recall bias in case-control studies of birth defects
Teratology
 , 
1994
, vol. 
49
 (pg. 
273
-
81
)
33.
Swan
SH
Shaw
GM
Schulman
J
Reporting and selection bias in case-control studies of congenital malformations
Epidemiology
 , 
1992
, vol. 
3
 (pg. 
356
-
63
)
34.
Martinez-Frias
ML
Rodriguez-Pinilla
E
Prieto
L
Prenatal exposure to salicylates and gastroschisis: a case-control study
Teratology
 , 
1997
, vol. 
56
 (pg. 
241
-
3
)
35.
Lam
PK
Torfs
CP
Brand
RJ
A low pregnancy body mass index is a risk factor for an offspring with gastroschisis
Epidemiology
 , 
1999
, vol. 
10
 (pg. 
717
-
21
)
36.
Kuhn
L
Kline
J
Ng
S
, et al.  . 
Cocaine use during pregnancy and intrauterine growth retardation: new insights based on maternal hair tests
Am J Epidemiol
 , 
2000
, vol. 
152
 (pg. 
112
-
19
)
37.
Loane
M
Dolk
H
Bradbury
I
, et al.  . 
Increasing prevalence of gastroschisis in Europe 1980 –2002: a phenomenon restricted to younger mothers?
Paediatr Perinat Epidemiol
 , 
2007
, vol. 
21
 (pg. 
363
-
9
)
38.
Forrester
MB
Merz
RD
Comparison of trends in gastroschisis and prenatal illicit drug use rates
J Toxicol Environ Health A
 , 
2006
, vol. 
69
 (pg. 
1253
-
9
)