RESEARCH REPORT

© 2003 Society for the Study of Addiction to Alcohol and Other Drugs

Addiction,

98

, 785–789

Blackwell Science, Ltd

Oxford, UK

ADDAddiction

1360-0443© 2003 Society for the Study of Addiction to Alcohol and Other Drugs

98Original Article

Maternal drug use and neonatal stay durationK. Johnson et al.

Correspondence to:

A. GreenoughDepartment of Child HealthKing’s College HospitalLondon SE5 9RSUKTel:

+

44 20 73463037Fax:

+

44 20 79249365E-mail: anne.greenough@kcl.ac.uk

Submitted 15 July 2002;initial review completed 24 October 2002;

final version accepted 3 February 2003

RESEARCH REPORT

Maternal drug use and length of neonatal unit stay

K. Johnson

1

, A. Greenough

1

& C. Gerada

2

Children Nationwide Regional Neonatal Intensive Care Centre, King’s College Hospital, London

1

and The Hurley Clinic, London, UK

2

ABSTRACT

Aims

Infants with neonatal abstinence syndrome (NAS) may require a pro-longed neonatal unit admission, which has implications for both their familiesand bed occupancy. The aim of this study was to test the hypothesis that theduration of neonatal unit stay would be influenced by the type of maternal druguse and particularly prolonged for the infants whose mothers had taken meth-adone with other substances.

Design

The medical records of infants born at term who were admitted consec-utively to a neonatal unit because of NAS were reviewed. Data were collectedregarding antenatal and neonatal factors likely to affect neonatal stay. Compar-isons were then made between three groups of infants: those whose motherstook methadone alone, methadone plus other drugs or non-methadone opioids.

Setting

Level three neonatal intensive care unit.

Participants

Forty-one infants with a median gestational age of 39 (range 37–42) weeks.

Findings

The 41 infants had a median duration of admission of 30 (range 3–68) days. Thirty-six of the infants required treatment for NAS; their medianduration of treatment was 29 (range 6–68) days. The duration of stay andrequirement for treatment were greater in the infants exposed to methadoneand other drugs compared to those exposed to non-methadone opiods only(

P

=

0.0212,

P

=

0.0343, respectively). The duration of stay without require-ment for treatment was also longest in the methadone plus other drugs group(

P

=

0.0117).

Conclusions

Prolonged treatment and neonatal unit stay are influenced by thetype of maternal drug abused.

KEYWORDS

Neonatal abstinence syndrome, neonatal intensive care.

INTRODUCTION

Infants with neonatal abstinence syndrome (NAS) canhave a prolonged admission to the neonatal unit. In onestudy, the mean duration of neonatal unit stay for aninfant with NAS was 22 days (Coghlan

et al

. 1999); thisimpacted unfavourably on bed occupancy, as evidencedby approximately three cots per day on the neonatal unitbeing required by affected infants. Such a prolongedadmission is also disruptive to the infant’s family. It istherefore important to determine which factors are asso-ciated with prolonged admission, as this would facilitateboth counselling of parents and identification of interven-tions likely to reduce stay.

There are a variety of factors which influence theduration of neonatal unit admission required by infantswith NAS. These include the development of severe NASand hence the need for prolonged treatment, antenataldrug exposure causing pregnancy complications result-ing in increased infant morbidity and adverse social cir-cumstances preventing early discharge. All those factorsmay be influenced by the type of drug(s) used duringpregnancy. More severe NAS is seen in infants of womenwho took methadone rather than diamorphine duringpregnancy (Kandall

et al

. 1977; Wilson

et al

. 1981). Pla-cental abruption is associated with cocaine abuse andantepartum haemorrhage with opiate abuse (Hulse

et al

.1998) and those complications could result in birth

© 2003 Society for the Study of Addiction to Alcohol and Other Drugs

Addiction,

98

, 785–789

786

K. Johnson

et al.

depression. If women have a chaotic life-style the infantmay have to remain in hospital until a suitable carer and/or appropriate discharge venue can be found. It is ourclinical experience that that adverse outcome is mostcommon in users of multiple illicit drugs, that is polydrugusers. It thus seems likely that the type(s) of drug usedantenatally would affect the length of admission and themost prolonged stays would be in infants whose motherstook methadone plus other drugs. The aim of this studywas to test those hypotheses.

METHODS

The outcomes of infants born at term and consecutivelyadmitted to the neonatal unit between 1991 and 2001because of NAS related to

in utero

opioid exposure werereviewed in 2002. During the study period infants withNAS, regardless of whether they were cared for on thepostnatal wards or on the neonatal unit, were assessed at4-hourly intervals using the River’s scoring system (Riv-ers 1986). Infants were assessed to determine whetherthey had any of 10 abnormal signs, scoring one for eachsign that was present. Infants were admitted to the neo-natal unit if they had a high score, that is more than threeon at least two consecutive occasions. Following admis-sion, urine was obtained for toxicology analysis to deter-mine which drugs the mother had been using. The urinewas screened for the presence of maternal drug abuse,which included screening for amphetamines, barbitu-rates, cannabis, benzodiazepines, cocaine, methadone,codeine, dihydrocodeine, heroin, morphine, opiatemetabolites and pethidine. If the infant’s score remainedhigh, treatment was given according to the neonatalunit’s protocol. In the first part of the study period (1991–96), infants with NAS were prescribed chlorpromazine;subsequently, morphine was used. The starting dose ofchlorpromazine was 3 mg/kg/day and of morphine was0.5 mg/kg/day. The type of treatment the infant receivedwas not modified according to the drugs that the motherused. Infants were scored 4-hourly; treatment was com-menced if, after admission to the unit, the infant scoredthree or more on two successive occasions. Therapy wasincreased or weaned according to the infant’s score, thesame policy was used for altering the dose of chlorprom-azine and morphine. While treatment was required theinfant remained on the neonatal unit. Infants were dis-charged from the neonatal unit only when all relevanthealth-care professionals agreed that the mother wasable to look after her infant at home or at another appro-priate venue or other suitable carers had been found.

The maternal notes were reviewed to determine thetype of antenatal drug use and whether the mother hadaccessed antenatal care (mothers were recorded as

receiving antenatal care if they had attended at least oneclinic) or had previous children in care. At the time ofthe study there was no dedicated antenatal clinic fordrug using women. From the infant’s records the follow-ing information was obtained: the gestational age atdelivery and birth weight, the timing of onset of with-drawal symptoms, whether treatment was required andthe duration of requirement for treatment and neonatalunit stay.

Analysis

The infants were divided into three groups according tothe type of drug to which they had been exposed

in utero

and the results of the urine analysis. The three groupswere methadone only (methadone group), methadoneplus other drugs, which were heroin (diamorphine),cocaine (crack or cocaine powder), benzodiazepines,amitryptyline, amphetamines and barbiturates (metha-done plus group) or non-methadone opiates only, whichwere heroin (diamorphine) (

n

=

9) and dihydrocodeine(

n

=

1) (non-methadone group). The Kolmogorov–Smirnov test demonstrated the data were not distributednormally. Differences between the groups were thereforeassessed for statistical significance using the

c

2

, Kruskal–Wallis or Mann–Whitney

U

-test, as appropriate.

Sample size

In a previous study the average duration of stay was22 days, but stays of up to 62 days were required(Coghlan

et al

. 1999). A sample size of 30 (at least 10infants into each group) would allow detection, with 90%power at the 5% level, to detect a difference of 22 days inthe length of stay between the groups

RESULTS

Women in the methadone and methadone plus groupswere significantly more likely to have received antenatalcare than those in the non-methadone group(

P

=

0.0123), but the proportion who had had a previouschild who was now ‘in care’ did not differ significantlybetween the three groups (Table 1). The median birthweight of the non-methadone group was significantlylower than the methadone group (

P

=

0.0121). Thirty-six of the 41 infants required treatment for their NAS; themedian duration of treatment was 28.5 (range 6–68)days. Neither the timing of onset of NAS nor the require-ment for treatment differed between the three groups(Table 2). The proportions of infants in the three groupswho received chlorpromazine or morphine did not differsignificantly. The median duration of treatment of the

Maternal drug use and neonatal stay duration

787

© 2003 Society for the Study of Addiction to Alcohol and Other Drugs

Addiction,

98

, 785–789

infants who received chlorpromazine was 10 (range 6–64) days and of those who received morphine was 28(range 6–68) days (non-significant). The duration ofadmission was longer in the methadone plus group com-pared to the other two groups, but only the comparisonwith the non-methadone group reached statistical signif-icance (

P

=

0.0212). Similarly, the treatment durationwas longer in the methadone plus group compared to theother two groups, but only the comparison with thenon-methadone group reached statistical significance(

P

=

0.0343). Both the duration of stay and treatmentrequirement was longer in the methadone compared tothe non-methadone group, but the differences were notstatistically significant. The duration of neonatal unitstay following cessation of treatment or if no treatmentwas required was significantly longer in the methadoneplus group when compared with the methadone alonegroup (

P

=

0.0117). The majority of infants were dis-charged home with their mother and this outcome didnot differ significantly between the groups.

DISCUSSION

We have demonstrated that infants whose mothers tookmethadone and illicit substances in pregnancy had sig-nificantly longer neonatal unit admissions than infants ofwomen with other types of drug use. Approximately 60%of the women took methadone and other drugs. In previ-ous studies (Wilson

et al

. 1981; Brown

et al

. 1998)(Coghlan

et al

. 1999) the proportion has been reported tobe as high as 80%. This form of drug use then has impor-tant implications for resource utilization. The majority ofwomen who take methadone have been enrolled in meth-adone prescribing programmes (Payot & Berner 2000)and therefore have had contact with health-care pro-fessionals. Such contact should offer the opportunityto counsel women regarding the adverse effects oftaking additional substances with methadone duringpregnancy.

To determine the type of drug use we undertook infantscreening as well as examining the maternal notes.

Table 1

Antenatal and neonatal outcomes. Data are expressed as median (range) or the number of patients (

n

).

Methadone alone Methadone plus Non-methadone

n

14 17 10Maternal age (years) 30 (21–37) 28 (24–37) 30 (21–37)Maternal smoking -No. of cigarettes

a

6–10 11–20 6–10No. of women having antenatal care 13 16 5No. of primigravid women 7 8 2Previous children in care 4 (7)

b

4 (9)

a

2 (6)

a

Subject discharged with mother 12 11 6Vaginal delivery 10 11 6Males 8 9 6Gestational age (weeks) 39 (37–42) 39 (37–39) 39 (37–40)Birth weight (kg) 2.91 (2.54–3.64) 2.92 (1.6–3.36) 2.54 (2.10–3.34)Head circumference (cm) 34 (32–36) 33 (31–36) 33 (31–34)

a

The number of cigarettes smoked was recorded as 0–5, 6–10, 11–20. Data were not available for seven mothers (five methadone plus).

b

Number in brackets denotes number of women with previous children.

Table 2

Treatment requirement and duration of stay. Data are demonstrated as median (range) or number of patients (

n

).

Methadone alone Methadone plus Non-methadone

n

14 17 10Age at symptom onset 1 (1–3) 2 (0–5) 1 (1–6) (days)Number of infants requiring 14 15 7treatmentNumber of infants receivingChlorpromazine 5 4 2Morphine 9 11 5Duration of treatment 25.5 (6–64) 37 (0–68) 8.5 (0–49) (days)Duration of stay without treatment (days) 2.5 (0–9) 5 (1–20) 4 (0–9)Total duration of stay 29 (9–64) 41 (12–68) 13 (3–49) (days)

© 2003 Society for the Study of Addiction to Alcohol and Other Drugs

Addiction,

98

, 785–789

788

K. Johnson

et al.

Maternal history alone is not a reliable method of deter-mining drug use. Screening by meconium analysis dem-onstrated that 44% of 3010 infants examined had apositive drug screen, yet only 11% of the mothers admit-ted to drug use (Ostrea

et al

. 1992). The infants in thisstudy underwent urine toxicology screening. Thismethod can generate false negatives, but this arises only ifthe mother does not use drugs for a few days before deliv-ery or it is not possible to obtain a urine sample from theinfant soon after birth (Ostrea

et al

. 1992).Women who take methadone rather than diamor-

phine have been reported to be more likely to access ante-natal care (Lifschitz

et al

. 1985; Edelin

et al

. 1988;Soepatmi 1994) and this has been associated with a bet-ter pregnancy outcome. Meta-analysis (Hulse

et al

. 1998)has demonstrated that methadone compared to diamor-phine use during pregnancy is associated with a lowerneonatal mortality. The majority of studies, however, arefrom the United States, where the pattern of care offeredfrequently differs from that provided in the Uniyed King-dom. We now demonstrate that women living in a UKinner city who took methadone, even if this was withillicit or non-prescribed substances, were significantlymore likely to access antenatal care than women usingnon-methadone opiates.

Methadone compared to diamorphine use has beenshown to have less effect in reducing birth weight. Meta-analysis (Hulse

et al

. 1997) demonstrated the meanreduction in birth weight associated with diamorphonewas 489 g compared to 279 g with methadone. In thepresent study those infants whose mothers took non-methadone opioids had lower birth weights than infantswhose mothers took methadone, regardless of whetherthis was with other substances. The combination ofdiamorphine with methadone has been shown by meta-analysis to have the greatest effect on birth weight reduc-tion (Hulse

et al

. 1997). That combination, however, wasunusual in this study. There are a number of explanationsfor the birth weight reduction in infants of drug usingwomen. All of the women for whom there were datasmoked and maternal smoking has been associated witha reduction in birth weight (Milner

et al

. 1999). Thewomen in the methadone plus group, however, tended tosmoke the most cigarettes, yet it was the non-methadonegroup’s infants who had the lowest birth weights. Inter-estingly, the latter group were the women who were leastlikely to access antenatal care.

Infants whose mothers had taken methadone plusother substances required significantly longer durationsof treatment for NAS and hence neonatal unit stay thaninfants in the other two groups. A prolonged treatmentduration has been reported previously in women whotake diazepam with methadone (Coghlan

et al

. 1999).The methadone only group also tended to require longer

treatment than those whose mothers took other opioidsduring pregnancy, confirming that NAS is more severe inwomen taking methadone. During the earlier part of thestudy period NAS infants were prescribed chlorprom-azine, and in the latter part morphine. The pattern ofdrug abuse did not change over the study period and thusthe type of NAS treatment did not influence our results.Chlorpromazine is a neuroleptic; early data suggestedthat chlorpromazine and phenobarbitone had similarefficacy in controlling the symptoms of NAS (Kahn

et al

.1969). Current evidence suggests that opioids are themost appropriate treatment for NAS suffered by infantsfrom

in utero

exposure to diamorphine or methadone an,as a consequence, we no longer use or recommend chlo-rpromazine to treat NAS. There are, however, very fewdata to inform the appropriate choice of treatment forinfants exposed to multiple drugs.

The majority of infants were discharged 1–2 days fol-lowing the cessation of treatment or were kept on the unitfor a similar period only if treatment for NAS was deemednot to be required. Nevertheless, some infants spent up to20 further days on the neonatal unit because the socialcircumstances of their mother prevented discharge. Oth-ers (Payot & Berner 2000) have also reported that thelength of neonatal unit stay may be increased by non-medical problems; in that study the average extension was8 days. To minimize neonatal unit stay, those data arguestrongly for pre-delivery planning meetings to ensureappropriate carer(s) and venue of care following dischargeare identified well before the infant is to be discharged.

In all three groups of infants the majority of the neo-natal unit stay was required because of treatment forNAS. The duration of stay could then be significantlyreduced by discharging home otherwise ‘healthy’ infantsonce they are stable on treatment. Drug-abusing parents,however, may not have sufficiently competent organiza-tional skills to comply with the follow-up needs of theirinfants. Indeed, in one study (Agarwal

et al

. 1999), only50% of drug abusing women kept their infant’s out-patient appointments. It has been suggested that compli-ance with clinical attendance can be improved by estab-lishment of a dedicated clinic, staffed by personnelalready well-known to the families and who provide forall the infant’s medical and immunization requirements(Oei

et al

. 2001). Provisions of such a facility has resourceimplications, and whether this is more cost–effective inimproving the long-term outcome of NAS infants thankeeping them on a neonatal unit until they no longerhave NAS treatment requirements requires appropriatetesting. In conclusion, length of neonatal stay is influ-enced by the type of maternal drug use. Accurate knowl-edge of the pattern of drug use in the antenatal clinic istherefore important for counselling prospective parentsand planning health service resources.

Maternal drug use and neonatal stay duration

789

© 2003 Society for the Study of Addiction to Alcohol and Other Drugs

Addiction,

98

, 785–789

REFERENCES

Agarwal, P., Rajadurai, V. S., Bhavani, S. & Tan, K. W. (1999)Perinatal drug abuse in KK Women’s and Children’s Hospital.

Annals of the Academy of Medicine, Singapore

,

28

, 795–799.Brown, H. L., Britton, K. A., Mahaffey, D., Brizendine, E., Heitt,

A. K. & Turnquest, M. A. (1998) Methadone maintenance inpregnancy: a reappraisal.

American Journal of Obstetrics andGynecology

,

179

, 459–463.Coghlan, D., Milner, M., Clarke, T., Lambert, I., McDermott, C.,

McNally, M., Beckett, M. & Matthews, T. (1999) Neonatalabstinence syndrome.

Irish Medical Journal

,

92

, 232–233.Edelin, K. C., Gurganious, L., Golar, K., Oellerich, D., Kyei-

Aboagye, K. & Adel Hamid, M. (1988) Methadone mainte-nance in pregnancy: consequences to care and outcome.

Obstetrics and Gynecology

,

71

, 399–404.Hulse, G. K., Milne, E., English, D. R. & Holman, C. D. J. (1997)

The relationship between maternal use of heroin and metha-done and infant birth weight.

Addiction

,

92

, 1571–1579.Hulse, G. K., Milne, E., English, D. R. & Holman, C. D. J. (1998)

Assessing the relationship between maternal opiate use andantepartum haemorrhage.

Addiction

,

93

, 1553–1558.Kahn, E., Neumann, L. & Polk, G. A. (1969) The course of heroin

withdrawl syndrome in newborn infants treated with phe-nobarbital or chlorpromazine.

Pediatrics

,

75

, 495–500.Kandall, S. R., Albin, S., Gartner, L. M., Lee, K. S., Eidelman, A. &

Lowinson, J. (1977) The narcotic-dependent mother: fetal

and neonatal consequences.

Early Human Development

,

1

,159–169.

Lifschitz, M. H., Wilson, G. S., O’Brian Smith, E. & Desmond, M.M. (1985) Factors affecting head growth and intellectualfunction in children of drug addicts.

Pediatrics

,

75

, 269–274.Milner, A. D., Marsh, M. J., Ingram, D. M., Fox, G. F. & Susiva, C.

(1999) Effect of smoking in pregnancy on neonatal lung func-tion.

Arch Dis Child Neonatal Ed

,

80

, F8–F14.Oei, J., Feller, J. M. & Lui, K. (2001) Coordinated outpatient care

of the narcotic-dependent infant.

Journal of Paediatrics andChild Health

,

37

, 266–270.Ostrea, E. M. Jr, Brady, M., Gause, S., Raymundo, A. L. & Stevens,

M. (1992) Drug screening of newborns by meconium analy-sis: a large-scale, prospective, epidemiologic study.

Pediatrics

,

89

, 107–113.Payot, A. & Berner, M. E. (2000) Hospital stay and short-term

follow-up of children of drug-abusing mothers born in anurban community hospital—a retrospective review.

EuropeanJournal of Pediatrics

,

159

, 679–683.Rivers, R. P. A. (1986) Neonatal opiate withdrawal.

Archives ofDisease in Childhood

,

61

, 1236–1239.Soepatmi, S. (1994) Developmental outcomes of children of

mothers dependent on heroin or heroin/methadone duringpregnancy.

Acta Paediatrica Supplement

,

404

, 36–39.Wilson, G. S., Desmond, M. M. & Wait, R. B. (1981) Follow-up of

methadone treated and untreated narcotic dependent womenand their infants: health developmental and social implica-tions.

Journal of Pediatrics

,

98

, 716–722.