Perinatal outcome of illicit substance use in pregnancy—comparative and contemporary socio-clinical profile in the UK
- First Online:
- Cite this article as:
- Goel, N., Beasley, D., Rajkumar, V. et al. Eur J Pediatr (2011) 170: 199. doi:10.1007/s00431-010-1284-6
- 563 Downloads
The aim of the study was to determine the contemporary socio-clinical profile and perinatal outcome of illicit substance use in pregnancy in a large UK city and compare with published literature. Cases were identified retrospectively from the ‘cause for concern’ referrals over 5 years (2003–2007). Data was collected on mother–infant pair from medical notes and laboratory records. Chi-square and Mann–Whitney U tests were used where appropriate for statistical analysis. One hundred sixty-eight women were identified as using illicit substance in pregnancy. Smoking (97.4%), unemployment (85.4%) and single status (42.3%) were frequent. Besides controlled use of methadone, heroin, cannabis and benzodiazepines were the most commonly used drugs. Hepatitis C prevalence was high (29.9%) despite low antenatal screening rates (57.7%). Neonatal morbidity was related to prematurity (22.9%), small for dates (28.6%) and neonatal abstinence syndrome (NAS; 58.9%). By day 5 of life, 95.1% of the babies developing NAS and 96.1% of those requiring pharmacological treatment were symptomatic. Of the infants developing NAS, 31.7% required pharmacological treatment. A total of 82.5% babies went home with their mother, and 21.2% were placed on the Child Protection Register. Only 14.3% were breast feeding at discharge. Illicit substance use in pregnancy continues to be associated with significant maternal and neonatal morbidity, and the socio-clinical profile in this decade appears unchanged in the UK. Hepatitis C prevalence is high, and detection should be improved through targeted antenatal screening. Where facility in the community is unavailable, 5 days of hospital stay is sufficient to safely identify babies at risk of developing NAS. Most babies were discharged home with their mother.
KeywordsIllicit substance use Substance misuse Socio-clinical profile Perinatal outcome Neonatal abstinence syndrome
Illicit substance use continues to be a major public health problem in the UK . Around 10% of adults aged 16–59 years, 22% aged 16–24 years and 8% aged 24–29 years reported using any illicit drugs in 2008/2009 in England and Wales . Substance-using women are mostly of childbearing age, and 18% of women between 16 and 24 years of age report drug use .
Substance use in pregnancy poses significant health risks for the mother, foetus and the newborn child and requires specialist care. Maternal risks are secondary to socioeconomic and lifestyle issues, infection-related morbidities, hazards related to the specific substance use and its procurement. Women constitute a quarter of the treatment population of substance use services, and over 90% are of childbearing age . These women are known to be poorly compliant and disengaged with healthcare resources, and the extent and ease with which such services are available are extremely variable across the country. The exact scope of the problem is unknown, as it relies much on voluntary disclosure .
The foetal risks are related to the effect of the substance and associated maternal lifestyle on the developing organs and placental function. Newborns are at risk of being born early and small, acquire perinatal infection and develop neonatal abstinence syndrome (NAS). The practice of prolonged hospitalisation, for identification and treatment of NAS, as well as to resolve any outstanding child protection issue, is variable in the UK. It is disruptive to the family with likely effects on bonding and establishment of breast feeding. A recent survey of UK neonatal units (NNU) showed that 29% of them discharged babies on medications to avoid prolonged hospitalisation .
Published data in the UK on perinatal outcomes of illicit substance use in pregnancy is sparse and mostly from very large cosmopolitan cities. Swansea is the second largest city in Wales and has a single large maternity unit that manages high-risk pregnancies. The catchment is a mixture of urban, semi-urban and rural population that has a higher unemployment and lower socioeconomic status than the Welsh average . The ‘Swansea Drugs Project’ is one of the oldest substance misuse agencies in Wales that works closely with the community drug and alcohol team.
The aim of this 5-year retrospective observational study was to determine the contemporary socio-clinical profile and perinatal outcome of illicit substance use in pregnancy in a large UK city. We focussed on specific maternal and neonatal outcomes including utilisation of hospital inpatient services and compared our findings with published data in the UK and from the rest of the world.
This retrospective observational study was undertaken between January 2003 and December 2007. The study was part of service evaluation and audit of standards of care in pregnancies with illicit substance use against local and national guidelines . The study was registered with the trust audit register, and the local research ethics committee has confirmed that there are no ethical issues related to publication of the results.
Pregnant women with history of illicit substance use were identified from the ‘Cause for Concern’ referral records at our hospital. Such referrals were usually made during the pregnancy following maternal disclosure to the midwife at booking or subsequently to other carers or agencies. Maternal disclosure was relied on to ascertain drug use rather than urinary drug screens, as the latter only provided a cross-sectional snapshot and was unlikely to detect all illicit drugs. Additionally, poor compliance in agreeing to the test was likely to underestimate the true prevalence. Women who had miscarriage or stillbirths or used prescribed medicines for pre-existing medical conditions were excluded. Women with previous history of drug use who abstained during this pregnancy were also excluded. For each mother–infant pair identified, a predefined dataset was completed using information from the medical case notes, laboratory records and other hospital databases.
Maternal demographic details, social history, profile of illicit substance use and information on antenatal care including results of infection screen were collected. Polydrug use was defined as use of more than one illicit drug excluding alcohol and tobacco.
Neonatal data included gestation, birth weight, mode of delivery and details of postnatal observations for detection of NAS. The hospital guideline for the management of NAS was extensively revised and implemented in 2007. Prior to this (2003–2006), infants born to substance-using mothers were required a mandatory hospital stay of 14 days to monitor symptoms of NAS. This was reduced to 7 days in 2007. Symptoms of NAS were recorded using the modified Lipsitz tool from 2003 to 2006 and modified Finnegan score in 2007. These two scoring systems, although relying heavily on observational measurements, are the most commonly used in the UK and the rest of the world [15, 17]. A persistent score of greater than 4 points on the Lipsitz scale and three scores of 8 points in a 24-h period on the Finnegan scale were generally regarded as threshold for pharmacological intervention [5, 12]. For the purpose of this study, infants scoring 2 points on the Lipsitz scale and 4 points on the Finnegan scale were regarded as ‘withdrawing’.
Reasons for admission to the NNU were identified. All babies requiring pharmacological treatment were admitted to the NNU. Co-morbidities including prematurity (<37 completed weeks of gestation), small for date (less than tenth centile), hypoglycaemia (<2.6 mmol/l on laboratory measurement or equivalent on point of care equipment), method of feeding and need for sepsis evaluation were recorded.
Babies of seropositive mothers (HBsAg ± HbeAg positive) and those at high risk of acquiring hepatitis B virus (HBV) infection, i.e. history of intravenous drug use, received hepatitis B vaccine ± immunoglobulin as appropriate. Follow-up data on HBV immunisation and seroconversion was obtained. Data was also collected on discharge destinations and enrolment on the Child Protection Register.
Anonymised data was analysed on Microsoft Excel 2003. In some categories, data was incomplete due to lack of documentation or unavailability of clinical notes. When reporting results, we have indicated the denominator in each category. Cohort characteristics were expressed as proportion, median and range as appropriate. Where appropriate, the non-parametric Mann–Whitney U test or Chi-square analysis of 2 × 2 contingency tables was used to determine the significance of outcome differences between independent groups of continuous or categorical variables, respectively. p < 0.05 was considered to be statistically significant.
Of the 17,856 pregnancies in this period, 168 (0.9%) were identified with history of illicit substance use in the current pregnancy.
Maternal age in years: median (range)
Heavy smoker (>10 cigarettes a day)
Without a current partner
Fifty-three of 64 (82.8%) known partners were also drug users. Exposure to domestic violence was seen in 17 out of 81 (21%) women where this information was specifically recorded. A previous child was in care in 29 cases. Of the 94 women receiving regular antenatal care, 22 (23.4%) were late bookers (>20 weeks).
Antenatal infection screening
Number screened, n/N
Small for gestation age (<tenth centile)
Hypoglycaemia (<2.6 mmol/l)
Resuscitation at birth
In 26/136 babies (19.1%), septic screen was performed to rule out infection, and antibiotics were given, pending blood culture reports. There were no positive cultures.
Eighty-two of 139 babies (58.9%), where records of withdrawal scores were available, developed withdrawal symptoms. A higher proportion of babies withdrew in the polydrug user group (56 of 81) as compared to monodrug users (26 of 51), and this difference was statistically significant (p = 0.03). Methadone was part of the mother's substance use in 54 out of the 82 babies (65%) with NAS. Of the 72 babies in the maternal methadone use group, 23 required pharmacological treatment as compared to 1 out of 28 in the non-methadone opiate use (p = 0.006). Of the 37 babies with maternal benzodiazepine use, 24 (64.9%) developed NAS, and seven (18.9%) required pharmacological treatment.
By day 5, 95.1% (78/82) babies who developed NAS were symptomatic, and 81.7% (67/82) had reached their peak symptoms. Pharmacologic treatment was required in 26/82 (31.7%) babies with NAS. All but one baby (96.1%) requiring pharmacological treatment developed NAS by day 5 of life. There was no significant difference between groups of polydrug and monodrug users regarding incidence of pharmacological intervention for NAS (20 vs. 6; p = 0.11). All except two babies were treated with morphine sulphate. The median duration of treatment was 13 days, and the median maximum dose of morphine was 240 mcg/kg/day. Chloral hydrate was used in conjunction with morphine in nine cases. The remaining two babies were treated with phenobarbitone and chlorpromazine, respectively.
Fifty-nine of 168 babies (35.1%) were admitted to NNU. Pharmacological treatment for NAS was the most frequent reason for admission to NNU (26/59; 44%). Other reasons for admission were prematurity, low birth weight, social issues and closer monitoring of NAS.
Seventy-one of the 130 babies born to HBV seronegative women (HBsAg negative) were offered hepatitis B vaccination due to associated risk factors. Only 25 (35.2%) completed the full course of four doses. In contrast, five of the seven babies (71.4%) born to HBsAg-positive mothers completed the full vaccination schedule and remained negative for HBV infection markers. Two of the seven HBsAg-positive mothers were also HbeAg positive, and their infants received hepatitis B immunoglobulin in addition to hepatitis B vaccine and completed the full immunisation schedule. Twenty-nine babies were born to mothers with positive serology for hepatitis C virus (HCV) infection. Full follow-up data was available in 23 babies, none of whom developed HCV infection.
Discharge outcomes were known for 137 babies, of which 113 (82.5%) were discharged home with their mother. Twenty-nine of 137 (21.2%) newborns were placed on the Child Protection Register, while 24/137 (17.5%) were placed in foster care. Of the 24 placed in foster care, 20 had pre-birth social services involvement, and the reasons for placement were a combination of domestic violence, unstable drug use in the family, concerns regarding parenting capabilities and history of previous children in care. Only 19/133 (14.3%) were breastfeeding at discharge.
The prevalence of illicit substance use in pregnant women was 0.9% (168/17,856) in our study. Recent data from the Northern Region in the UK and Dumfries in Scotland 2002 suggested an incidence of 0.75% and 0.11%, respectively [13, 18]. In a Croatian study (1997–2007), Vucinovic et al. reported a prevalence of 0.2% . Much higher prevalence has been reported from the US and Australia [11, 15].
Comparative data on reported maternal and neonatal outcomes of substance use in pregnancy in the UK (1997–2007)
Hull 1997–2003 
Dumfries 2001–2005 
Median 25 years
91%, 42.3% single
Heroin, methadone, cannabis
Methadone, heroin, cannabis
40% of screened
35% of screened
29.9% of screened
4% of screened
5.1% of screened
Besides controlled use of methadone, heroin and cannabis were the most frequent drugs used in pregnancy, similar to other reports from the UK [7, 13, 18]. In contrast, in the UK general population, cannabis and cocaine are the most commonly used drugs . Heroin use is known to be high in a population like ours with low socioeconomic status. Cannabis use may have been underreported by women in view of its minimal effect on neonatal abstinence. Polydrug use was common, despite an active methadone programme. This trend is similar to that noted in other studies [1, 2, 9, 24].
The lifestyle associated with drug use increases vulnerability to infection. Our study cohort had no HIV infection, a low prevalence of HBV but a relatively high HCV infection. The prevalence of HBV, HCV and HIV infection among intravenous drug users in the UK were 21%, 50% and 1%, respectively, with geographic variations . Analysis of paired HCV and HIV prevalence data has shown that HCV prevalence of up to 30% is associated with zero or very low prevalence of HIV . The prevalence of HCV infection had risen from our own historic data (5% in 1997–2002, unpublished) but matched closely with reports from other parts of the UK [7, 18]. Higher prevalence has been reported in Europe and by studies examining specifically intravenous drug users [6, 10, 13]. However, of major concern is the low HCV screening rate coupled with its high prevalence. Screening for HCV is not part of the routine antenatal screening in the UK. Treatment for HCV infection is effective and widely available. There is a need to improve awareness, counselling and screening rates of HCV infection in this population.
Prematurity (21.9%) was common in our cohort, much higher than the reported 6% in the general population . Although more than a quarter of the infants were SFD, the mean birth weight of 2.76 kg was similar to other reports [7, 9, 10, 21]. The higher proportion of non-instrumental deliveries (82% vs. 60–70% in Welsh population) could be explained by the small size of the babies secondary to prematurity and SFD . Despite a lack of culture-proven sepsis, many sepsis evaluations were undertaken due to overlap of symptoms of NAS and early onset sepsis.
The incidence of NAS (58.9%) was similar to that reported elsewhere [5, 12, 21]. The higher incidence of NAS with maternal polydrug use could be explained by greater opiate use in this group as compared to those using a single drug (91.35% vs. 58.8%). Methadone use, implicated in 65% of our cohort with NAS, has reported association with prolonged and severe NAS [5, 9, 15, 21].
Differences in admission policy, infrastructure for treatment of NAS (postnatal ward/community) and the efficiency with which safeguarding issues are managed influence both NNU admission rate and duration of hospital stay. Pharmacological treatment of NAS on the postnatal ward as compared to NNU has been shown to reduce total duration of hospital stay . A study of practices of NAS management in the UK found inconsistent policies in hospitals for its detection and treatment . Our organisation did not have the facility to institute pharmacological treatment on the postnatal ward or in the community. However, mothers had full access to their children at all times and took part in routine postnatal care. Despite a falling trend, greater than a third of the at-risk newborns required admission to NNU. The fall in NNU admission could be attributed to better recognition of NAS severity and institution of non-pharmacological treatment on the postnatal ward. The duration of hospital stay in our cohort was shorter than some reports, but similar to that reported from Hull [2, 7, 15]. The median hospital stay reduced following introduction of new guidelines without a concomitant increase in readmission rates.
The use of morphine sulphate in our study for treatment of NAS was similar to other reports from UK and elsewhere [15, 17, 20]. The median duration of pharmacological treatment was lower than that previously reported . The reasons are unclear but may be secondary to better antenatal control and consistent treatment regimens.
Our finding that 95.1% of the babies developing NAS and, more crucially, 96.1% requiring pharmacological treatment were symptomatic by day 5 of life is highly significant. We recommend that health infrastructures that do not have facilities for detection and management of NAS in the community could safely reduce the duration of in-hospital stay to identify infants at risk of NAS to 5 days thereby freeing up acute hospital beds and accompanying resources. Previous reports had also suggested that delayed presentation of NAS is uncommon even in methadone users . Oei et al. had suggested a similar plan of action, but only through coordinated outpatient care [14, 15].
We had no vertical transmission of HCV or HBV. Despite appropriate initiation of hepatitis B vaccination, compliance with the full vaccination schedule was poor. The poor compliance is likely due to asymptomatic status, lack of awareness and frequent change of abode in this cohort. Compliance was worse in infants of HBV seronegative mothers.
Despite active encouragement for breast feeding, only 14.3% babies were breastfed at discharge as compared to 50% in our general population. Four out of five babies were discharged home with their mother, similar to reported trends in UK and elsewhere [5, 7, 10]. We do not have data to determine if this pattern was sustained in the long term. Majority of foster placements were determined by pre-birth conferences and due to multiple social factors.
Our study has limitations. As a retrospective study, we had to rely on documentations in the case notes. There were missing data in some categories, but they constituted only a small proportion of the results reported. We are therefore confident that our data is representative of our population. We did not collect data on miscarriages and stillbirths. Majority of such events occurred even before a disclosure or ‘cause for concern’ was raised and therefore would have led to serious underestimates of such outcome. We did not have specific information on intravenous drug use or long-term follow-up data on most outcomes beyond hospital discharge. Future studies should address these deficiencies.
Limitations withstanding, our paper highlights important and contemporary socio-clinical profile of substance use in pregnancy and its effect on maternal and neonatal morbidity. These are likely to be important in planning and delivery of healthcare to this vulnerable group.
Illicit substance use in pregnancy continues to be associated with significant maternal and neonatal morbidity and a vulnerable social background. The socio-clinical profile over the last decade appears similar in reported studies across urban areas in the UK. Hepatitis C prevalence is high, and detection through targeted antenatal screening should be improved. Where monitoring facilities are not available in the community, 5 days of hospital stay is sufficient to safely identify babies at risk of developing NAS. Majority of the babies are discharged home with their mother (Appendix).
We would like to thank Virginia Hewitt, lead midwife for vulnerable adults and children, for her suggestions in the formative stage of the study and Cerys Nicholls for her help in collecting maternal data. We would like to thank the audit department at Singleton Hospital and Mrs. Lynda Challacombe for their help in obtaining medical notes.
Conflict of interest
The authors declare no conflict of interest.