Delivery timing for the opioid–exposed infant

Abstract

BACKGROUND:

The prevalence of opioid use disorder and medication-assisted treatment in pregnancy is increasing. Compared with term infants, preterm infants have a lower incidence of neonatal opioid withdrawal syndrome. It is unknown whether early term delivery compared with full or late-term delivery decreases the risk of neonatal opioid withdrawal syndrome.

OBJECTIVE:

This study aimed to compare the neonatal outcomes among opioid–exposed infants born in the early, full, and late-term periods.

STUDY DESIGN:

This was a retrospective cohort study of opioid–exposed pregnancies delivering at a single center from 2010 to 2017 at ≥37 weeks gestation. Participants with multiple gestations or fetal anomalies were excluded. Maternal opioid exposure was defined as prescription (including medication-assisted treatment) or nonprescription opioid use or a positive urine drug screen in pregnancy for opiates. The primary outcome was a neonatal composite of respiratory distress syndrome, neonatal sepsis, neonatal seizures, hypoxic ischemic encephalopathy, jaundice requiring treatment, 5-minute Apgar <5, neonatal intensive care unit admission, neonatal opioid withdrawal syndrome, or neonatal death. The secondary outcomes included individual components of the primary outcome, birthweight, need for and length of neonatal opioid withdrawal syndrome treatment, length of hospital admission, and maximum Finnegan scores. Early (37–<39), full (39–<41), and late (41–<42 weeks) term groups were defined by the American College of Obstetricians and Gynecologists.

RESULTS:

Of 399 infants, 136 (34.1%), 229 (57.4%), and 34 (8.5%) were born in the early, full, and late-term periods, respectively. Two hundred and seventy patients (67.7%) received medication-assisted treatment for opioid use disorder, and the baseline characteristics were similar in all the groups except for history of intranasal heroin use, positive urine toxicology screen for heroin or any opiates, and delivery indication (P<.05). The primary composite outcome occurred in 313 (78.4%) neonates, and 296 (74.2%) neonates had neonatal opioid withdrawal syndrome. More than half (219 [54.9%]) of opioid–exposed neonates were admitted to the neonatal intensive care unit, and 160 (40.1%) required pharmacologic neonatal opioid withdrawal syndrome treatment for a mean duration of almost 3 weeks (19.0±16.1 days). There were no significant differences in the primary composite outcome, incidence of neonatal opioid withdrawal syndrome, or other secondary outcomes (except birthweight) between neonates born in the early, full, or late-term periods.

CONCLUSION:

Although neonatal morbidity was frequent among opioid–exposed neonates, the incidence and severity of neonatal opioid withdrawal syndrome or other neonatal outcomes were not different between neonates delivered in the early, full, and late-term periods, suggesting that opioid–exposed infants may not benefit from early term delivery.

The opioid use epidemic in pregnancy parallels that in the general population. Approximately 15% of pregnant people have a substance use disorder, and opioid-related diagnoses in pregnancy have increased by 131% between 2010 and 2017.¹ Further, more than 20% of pregnant persons reported prescription opioid misuse in a self-reported opioid use survey.² Medication-assisted treatment (MAT) or opioid agonist pharmacotherapy is the mainstay of treatment owing to lower relapse and overdose rates. However, various studies have documented the association between any opioid use, including MAT, and neonatal opioid withdrawal syndrome (NOWS).^3-7 Between 2000 and 2009, the incidence of NOWS increased from 1.2 to 5.8 per 1000 hospital births.^8,9 More recently, between 2010 and 2017, an 82% increase in the incidence of NOWS was reported along with longer newborn hospital stays and associated 8-fold higher healthcare costs.^10-13

Previous studies have reported a lower incidence and severity of NOWS in preterm neonates compared with term neonates,^14-17 possibly owing to placental immaturity (with lower rates of fetal opioid exposure), immature hepatic metabolism of opioids in premature neonates, or immaturity of the central nervous system.^2,4 This led to the suggestion that opioid–exposed infants may benefit from early term delivery to decrease the risk of NOWS.¹⁸ However, among term neonates, whether the incidence of NOWS is less frequent in early term neonates remains unknown.

Therefore, our study aimed to examine opioid-related neonatal morbidity in opioid–exposed term neonates categorized by the American College of Obstetrics and Gynecology term definitions as follows: early (37–<39 weeks), full (39–<41 weeks), and late (41–<42 weeks) term. We hypothesized that earlier timing of delivery in opioid–exposed term neonates would be associated with improved opioid-related neonatal outcomes, including a reduced incidence of NOWS.

Materials and Methods

This was a retrospective cohort study of patients delivering at a single US tertiary center from January 2010 to December 2017. Included patients were pregnant people delivering at ≥37 weeks’ gestation with a history of opioid exposure during the current pregnancy, including intravenous or intranasal heroin and nonprescribed and prescribed opioids; patients with no reported history of opioid use but with a positive urine drug screen at any point during pregnancy were also included. Patients receiving MAT with buprenorphine, buprenorphine-naloxone, or methadone were also included. Patients with multiple gestations or fetal anomalies were excluded. Approval for this study was granted by the institutional review board of The University of Alabama at Birmingham.

The primary exposure was delivery period at term (≥37 weeks’ gestation) divided into 3 groups as defined by American College of Obstetricians and Gynecologists: early term (37–<39 weeks’ gestation), term (39–<41 weeks’ gestation), and late term (41–<42 weeks’ gestation).¹⁹ The primary outcome was a neonatal composite of respiratory distress syndrome (defined on the basis of clinical features and oxygen or respiratory support for ≥6 hours within 24 hours after birth), suspected neonatal sepsis, neonatal seizures, hypoxic ischemic encephalopathy (defined as clinical findings consistent with abnormal consciousness, depressed tone or reflexes, and abnormal respiration or seizures and evidence of asphyxia with an arterial cord gas pH<7.0), jaundice requiring treatment, 5-minute Apgar<5, neonatal intensive care unit admission, NOWS, or neonatal death. At our institution, infants are admitted to the neonatal intensive care unit (NICU) if born at <2 kg, for hypoglycemia <40 mg/dL refractory to enteral nutrition, or in case of respiratory distress as determined by the clinician. Similarly, all infants with known or suspected opioid exposure in-utero are observed following delivery for at least 5 days. Infants are typically roomed in with their parent(s) until parental discharge, which then necessitates neonatal transfer to the continuing care nursery–a unit that cares for neonates of ≥24 weeks gestation who do not require ventilator care.

NOWS was defined using the pharmacologic threshold; at our institution, NOWS is said to occur on having 2 consecutive Finnegan scores ≥12 or 3 consecutive scores ≥8 as assessed by the pediatrician at birth and every 3 to 4 hours throughout the hospitalization. Before pharmacotherapy initiation, the systematically employed nonpharmacologic interventions include skin-to-skin contact, minimizing external stimulation; on-demand feeding; and other previously described interventions.²⁰

The secondary neonatal outcomes were individual components of the primary outcome, birthweight, need for and length of NOWS treatment, length of hospital admission, and maximum Finnegan scores. Race and ethnicity were self-reported by the study participants into prespecified categories. However, neither race nor ethnicity was included in our analyses as covariates, as there have not been sufficient associations that may reflect differences in outcomes of opioid–exposed infants by race rather than underlying social determinants of health.

The baseline characteristics and outcomes were compared between the gestational age groups using analysis of variance or Kruskal–Wallis test for continuous variables and Pearson chi-square test or complementary exact tests for categorical variables. The trends in binary outcomes were evaluated with the Cochran–Armitage test or exact Cochran–Armitage test. We estimated that a sample size of 249 (83 for each group) would provide 80% power to detect a trend at rates of 0.8, 0.7, and 0.6 in the primary composite outcome across all gestational age groups at a type 1 error rate of 5%, assuming that all 3 groups are of equal size.

All the analyses were conducted with SAS 9.4 (SAS Institute Inc, Cary, NC), and sample size estimation was done with R version 4.1.3 (R Core Team). Significance level was set at a P value of <.05 for all analyses with no adjustment for multiple comparisons.

Results

Of 399 eligible participants, 136 (34.1%), 229 (57.4%), and 34 (8.5%) were born in the early, full, and late-term periods, respectively. Ninety (22.6%) were nulliparous, 171 (42.9%) were admitted in spontaneous labor, and 273 (68.4%) had a vaginal delivery. Two hundred and seventy (67.7%) participants were on MAT for opioid use disorder with buprenorphine, buprenorphine-naloxone, or methadone. The baseline characteristics were not significantly different between patients in the early, full, and late-term groups, except for history of intranasal heroin use, positive urine toxicology screen for heroin or any opiates, and delivery indication (P>.05). Other characteristics are shown in Table 1.

Table 1.

Baseline characteristics of opioid–exposed pregnancies by delivery period at term

Characteristic		Early terma (n=136)		Full terma (n=229)		Late terma (n=34)		P value
Maternal age at delivery (y)		29.3±5.1		28.8±4.6		28.6±4.3		.57
Race								.31
Black		21 (15.4)		23 (10.0)		2 (5.9)
White		111 (81.6)		202 (88.2)		32 (94.1)
Asian		1 (0.7)		0 (0.0)		0 (0.0)
None of the above		2 (1.5)		2 (0.9)		0 (0.0)
Unknown		1 (0.7)		2 (0.9)		0 (0.0)
Ethnicity								.49
Hispanic		4 (2.9)		3 (1.3)		0 (0.0)
Non-Hispanic		124 (91.2)		209 (91.3)		33 (97.1)
Unknown		8 (5.9)		17 (7.4)		1 (2.9)
Nulliparous		30 (22.1)		50 (21.8)		10 (29.4)		.61
Body mass index		29.9±7.2		30.7±6.5		31.8±5.7		.30
Married		27 (19.9)		47 (20.5)		6 (17.6)		.78
Medicaid health insurance		120 (88.2)		205 (89.5)		32 (94.1)		.92
Delivery indication								<.01
Spontaneous labor		60 (44.1)		108 (47.2)		3 (8.8)
Prelabor rupture of membranes		3 (2.3)		5 (2.2)		0 (0.0)
Oligohydramnios		2 (1.5)		4 (1.7)		0 (0.0)
Nonreassuring antenatal fetal testing		9 (6.8)		11 (4.8)		2 (5.9)
Preeclampsia		16 (12.0)		5 (2.2)		1 (2.9)
Other maternal comorbiditiesb		40 (30.1)		94 (41.0)		28 (82.4)
Vaginal delivery		90 (67.7)		158 (69.0)		25 (73.5)		.58
Opioid use historyc
Intravenous heroin		64 (47.1)		100 (43.7)		15 (44.1)		.82
Intranasal heroin		6 (4.4)		3 (1.3)		3 (8.8)		.03
Nonprescribed opioids		71 (52.2)		119 (52.0)		21 (61.8)		.55
Prescribed opioids		82 (60.3)		153 (66.8)		24 (70.6)		.35
No history with positive urine drug screen		16 (11.8)		14 (6.1)		1 (2.9)		.11
Urine toxicologyc
Amphetamines	13 (9.6)		21 (9.2)		4 (11.8)		.89
Barbiturates	6 (4.4)		10 (4.4)		0 (0.0)		.46
Benzodiazepines	43 (31.6)		53 (23.1)		6 (17.7)		.11
Buprenorphine	24 (17.7)		55 (24.0)		11 (32.4)		.13
Cannabis	24 (17.7)		32 (14.0)		2 (5.9)		.21
Cocaine	18 (13.2)		22 (9.6)		3 (8.8)		.52
Heroin	13 (9.6)		7 (3.1)		0 (0.0)		.01
Methadone	33 (24.3)		40 (17.5)		7 (20.6)		.29
Oxycodone	9 (6.6)		6 (2.6)		1 (2.9)		.16
Hydrocodone	2 (1.5)		2 (0.9)		0 (0.0)		.74
Any Opiates	63 (46.3)		73 (31.9)		9 (26.5)		.01
Medication-assisted treatment	84 (61.8)		158 (69.0)		28 (82.4)		.11
Tobacco use							.43
Never smoked	20 (14.7)		25 (10.9)		3 (8.8)
Current smoker	94 (69.1)		178 (77.7)		29 (85.3)
Former smoker	16 (11.8)		19 (8.3)		2 (5.9)
Maternal comorbiditiesc
Psychiatricd	63 (46.3)		91 (39.7)		11 (32.4)		.25
Hepatitis C infection	52 (38.2)		83 (36.2)		7 (20.6)		.13
Chronic hypertension	17 (12.5)		18 (7.9)		2 (5.9)		.23
Diabetes mellitus	6 (4.4)		13 (5.7)		2 (5.9)		.88
Method of feeding							.17
Breastfeeding	21 (15.4)		56 (24.5)		9 (26.5)
Formula feeding	77 (56.6)		104 (45.4)		16 (47.1)
Both breast and formula feeding	30 (22.1)		50 (21.8)		5 (14.7)

Data are presented as number (percentage) or mean±standard deviation, as appropriate.

^aEarly term: 37.0–38.6 weeks; full term: 39.0–40.6 weeks; late term 41.0–41.6 weeks.

^bOther maternal comorbidities included previous uterine surgery, placental previa, stillbirth, and elective labor induction.

^cPercentages sum up to more than 100 owing to overlapping subsets.

^dPsychiatric comorbidities include bipolar disorder, major depression disorder, and schizophrenia.

The primary composite outcome occurred in 313 (78.4%), neonates and 296 (74.2%) neonates had NOWS. The incidence of NOWS by week of gestation is shown in the supplemental figure. One hundred and sixty neonates (40.1%) required pharmacologic treatment for NOWS for a mean duration of treatment of 19.0±16.1 days. Two hundred and nineteen (54.9%) neonates were admitted to the NICU, and 216 (54.1%) were eventually admitted to the continuing care nursery. Other secondary outcomes are shown in Table 2. There were no significant differences in the rate of the primary outcome or incidence of NOWS between neonates delivered in the early, full, or late-term periods. Further, apart from birthweight, there were no significant differences in the secondary outcomes by gestational age groups at delivery (Table 2) (P trend >.05).

Table 2.

Primary and secondary outcomes of opioid–exposed pregnancies by delivery period at term

Outcomes	Early term (n=136)	Term (n=229)	Late term (n=34)	P value
Primary composite outcome	109 (80.1)	179 (78.2)	25 (73.5)	.44
NOWS	97 (71.3)	175 (76.4)	24 (70.6)	.86
NOWS pharmacologic treatment	58 (42.6)	89 (38.9)	13 (38.2)	.29
Mean maximum Finnegan score	12.1±5.3	12.1±4.9	11.5±5.5	.83
Length of NOWS treatment	16.2±12.6	20.4±18.4	21.9±12.5	.24
NICU or CCN admission	91 (66.9)	151 (65.9)	21 (61.8)	.42
Inpatient length of stay (d)	17.2±13.2	18.6±18.0	21.7±19.6	.51
Birthweight	2807.2±409.0	3217.8±420.3	3350.0±361.7	<.001
Respiratory distress syndrome	14 (10.3)	13 (5.7)	3 (8.8)	.25
Neonatal sepsis	16 (11.8)	20 (8.7)	2 (5.9)	.22
Neonatal jaundice requiring treatment	14 (10.3)	16 (7.0)	2 (5.9)	.27
Neonatal seizures	0 (0.0)	1 (0.4)	0 (0.0)	1.0
Hypoxic ischemic encephalopathy	0 (0.0)	1 (0.4)	0 (0.0)	1.0
5-min APGAR <5	0 (0.0)	2 (0.9)	0 (0.0)	.55
Neonatal death	0 (0.0)	0 (0.0)	0 (0.0)	—

Data are presented as number (percentage) or mean±standard deviation, as appropriate.

CCN, continuing care nursery; NICU, neonatal intensive care unit; NOWS, neonatal opioid withdrawal syndrome.

Comment

Principal findings

In our single-center cohort, adverse neonatal complications and NOWS occurred in more than 7 in 10 opioid–exposed neonates, and >40% required pharmacologic treatment with methadone for NOWS. In addition, more than half of opioid–exposed neonates required admission to a special care unit after birth. Apart from a lower mean birthweight in neonates delivered in the earlier term periods, timing of delivery was not significantly associated with either the primary composite outcome or any of the secondary neonatal outcomes. Of note, during the study period, our institution did not have an early term delivery policy for opioid–exposed pregnancies.

Results in context of what is known

The high rate of NOWS in our cohort has been previously reported in other populations. In a cohort of 18,869 patients with prepregnancy chronic opioid use, the rates of NOWS were 0.8%–1.7% depending on frequency and dose of opioid. Importantly, there were no patients with a history of opioid use disorder or on MAT in this cohort.²¹ Another cohort study of 872 pregnant patients on opioid maintenance therapy with buprenorphine or methadone found that 415 (58%) neonates had NOWS.²² Although some of these differences could be attributed to the type of opioid exposure (prescription opioids vs MAT) and ascertainment of NOWS from International Classification of Diseases codes in previous studies, it likely reflects population differences in the dose and duration of opioid exposure. Further, less than half of our population were breastfeeding, which is protective against NOWS.²³

In a study comparing late preterm (34°–<37 weeks’ gestation) and term (≥37 weeks’ gestation) neonates, there was a significant association between earlier gestational age at delivery and reduced need for NOWS pharmacotherapy.²⁴ Another study found that the need for pharmacotherapy for NOWS in preterm, term, and late-term infants was the same, but longer treatment courses were needed for term neonates (late preterm median, 16.0 [interquartile range (IQR), 10.0–24.0], early term, 22.5 [IQR, 15.0–40.0], full term, 23.0 [IQR, 6.0–38.0], and late term, 22.0 [IQR, 6.0–28.0] days, P=.02).¹⁷ This was similar to our findings that neonates delivered in earlier gestational periods had a shorter mean duration of NOWS treatment (early term 16.4±12.6 days, full term 20.4±18.4 days, and late term 21.9±12.5 days). In our population, these differences were not statistically significant, though we are not powered to detect this difference. Taken together, we postulate that placental and nervous system immaturity may not play a significant role in opioid-related neonatal complications among opioid–exposed neonates across the spectrum of term birth.²⁵

Clinical implications

Delivery timing in opioid–exposed term neonates was not associated with changes in opioid-related adverse neonatal complications. Given the known perinatal risks of early term deliveries compared with full term deliveries,²⁶ in the absence of other medical indications for delivery, term opioid–exposed infants do not appear to benefit from planned early term delivery.

Even though our study found high rates of NOWS in pregnant patients using opioids, these data should not be used to discourage adequate maternal treatment of medical conditions requiring opioid therapy, such as opioid use disorder, sickle cell disease, and others.²⁷ In Alabama, substance use disorder contributes to nearly half of maternal deaths²⁸—an emerging trend nationwide.²⁹ As MAT is associated with a lower rate of postpartum relapse and overdose, we also strongly encourage it, as the benefits to the maternal-neonatal dyad outweigh the risks of untreated opioid use disorder.

Research implications

Future studies should focus on identifying the factors that determine the severity of NOWS in opioid–exposed neonates. Furthermore, whether the risk of NOWS is significantly higher in patients receiving MAT than in those receiving other prescription or nonprescription opioids remains unclear. Future studies to examine these differences may be useful in counseling patients regarding the anticipated neonatal course. Finally, the specific MAT dose at which the risk of NOWS is significantly increased in term neonates remains undetermined.

Strengths and limitations

The strengths of our study include the use of a well-characterized cohort with rigorous opioid exposure, including data on dose and duration of MAT treatment, and neonatal outcome ascertainment over 8 years to evaluate differences by delivery gestational period. Further, the diagnosis of NOWS was made by a group of neonatologists on the basis of a standardized protocol. Although limitations of the Finnegan score system have been previously described, it remains one of the most widely used systems for NOWS assessment, improving the generalizability of our findings and involving other populations where NOWS is assessed with the Finnegan scale.^30,31 Last, our study was also powered to detect a 20% difference across the 3 groups in the primary composite across all the groups.

Our study is not without limitations. Data were obtained retrospectively through chart review, which is prone to ascertainment bias. The data cover an 8-year time period, during which there were changes in the clinical guidelines for the treatment of NOWS in opioid-exposed infants. Although these may have affected the duration and location of treatment, it is unlikely to have had any significant effect on other outcomes. Moreover, the distribution of comorbidities and the availability of social services that enhance MAT adherence and prenatal care uptake in our population likely differ from other centers, which may limit generalizability. Despite urine drug testing during prenatal care visits as indicated (usually before refilling prescriptions for MAT), we could not quantify the number of patients who, in addition to MAT, were also exposed to nonprescription opioids. Further, though MAT adherence is rigorously assessed at each prenatal visit, compliance to prescribed doses could not be ascertained. We also could not assess the effect of polysubstance use on our diagnosis and management, as up to 25% of our study population also had a positive urine toxicology screen for benzodiazepines at delivery. The number of late-term infants fell below our calculated sample size, and in an unbalanced design like ours, our final sample size requirements would increase to 477 (163, 274, 40) to achieve 80% power to detect the hypothesized trend. Lastly, we were not powered to detect differences by narrower week intervals, which may be more clinically meaningful to patients in planning a scheduled labor induction at a particular week of gestation to minimize the neonatal risks of opioid exposure.

Conclusions

Neonatal adverse outcomes and NOWS are not significantly different between opioid–exposed infants born in the early, full and late-term periods, suggesting that early term delivery may not benefit opioid–exposed infants. This study provides preliminary data for future studies examining the timing of delivery in pregnant people with opioid use disorder.

AJOG MFM at a Glance.

Why was this study conducted?

The incidence of maternal opioid use and neonatal withdrawal syndrome (NOWS) is rising. Compared with infants born at term, preterm neonates have a lower incidence of NOWS. However, among term infants, it is unclear whether gestational age at delivery in opioid–exposed neonates, is associated with the incidence of NOWS and other opioid-–related neonatal complications.

Key findings

NOWS occurs in more than 70% of opioid–exposed term neonates. The timing of delivery at term is not associated with an adverse primary composite neonatal outcome or the incidence of NOWS and need for treatment.

What does this add to what is known?

Early term delivery compared with full and late-term delivery of opioid–exposed neonates does not reduce the risk of opioid–associated neonatal complications.