Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Feb 10.
Published in final edited form as: Drug Saf. 2021 Oct 5;44(12):1297–1309. doi: 10.1007/s40264-021-01115-6

Maternal Complications and Prescription Opioid Exposure During Pregnancy: Using Marginal Structural Models

Xuerong Wen 1, Shuang Wang 1, Adam K Lewkowitz 2, Kristina E Ward 3, Erin Christine Brousseau 2, Kimford J Meador 4
PMCID: PMC8830421  NIHMSID: NIHMS1771419  PMID: 34609720

Abstract

Introduction

Prescription opioids are frequently used for pain management in pregnancy. Studies examining perinatal complications in mothers who received prescription opioids during pregnancy are still limited.

Objectives

The aim of this study was to assess the association of prescription opioid use and maternal pregnancy and obstetric complications.

Methods

This retrospective cohort study with the Rhode Island (RI) Medicaid claims data linked to vital statistics throughout 2008–2015 included pregnant women aged 12–55 years with one or multiple live births. Women were excluded if they had cancer, opioid use disorder, or opioid dispensing prior to but not during pregnancy. Main outcomes included adverse pregnancy and obstetric complications. Marginal Structural Cox Models with time-varying exposure and covariates were applied to control for baseline and time-varying covariates. Analyses were conducted for outcomes that occurred 1 week after opioid exposure (primary) or within the same week as exposure (secondary). Sensitivity studies were conducted to assess the effects of different doses and individual opioids.

Results

Of 9823 eligible mothers, 545 (5.5%) filled one or more prescription opioid during pregnancy. Compared with those unexposed, no significant risk was observed in primary analyses, while in secondary analyses opioid-exposed mothers were associated with an increased risk of cesarean antepartum depression (HR 3.19; 95% CI 1.22–8.33), and cardiac events (HR 9.44; 95% CI 1.19–74.83). In sensitivity analyses, results are more prominent in high dose exposure and are consistent for individual opioids.

Conclusions

Prescription opioid use during pregnancy is associated with an increased risk of maternal complications.

1. Introduction

Opioid analgesia has been commonly prescribed for pregnant women although the American College of Obstetricians and Gynecologists (ACOG) has recommended alternative pain treatments during pregnancy [1–5]. About one in seven women enrolled in private insurance plans and one in five Medicaid enrolled women are dispensed opioid analgesia during pregnancy [3, 4]. Current research on opioid use in pregnancy mostly focuses on the risk of adverse neonatal outcomes and congenital malformations [6–10]. Few studies have addressed perinatal complications in mothers who received prescription opioids during pregnancy. Pregnant women with opioid exposure may experience higher rates of anxiety, depression, acute cardiac events, chronic medical conditions, and maternal deaths from opioid overdose [10, 11]. A recent European study suggested a potential causal relationship between the genetic liability for increased use of prescription opioids and the risk of depression and anxiety [12]. Previous studies suggested a significant impact of maternal opioid use during pregnancy on perinatal morbidity and mortality, including poor fetal growth (adjusted odds ratio [aOR] 1.6, 95% confidence interval [CI] 1.5–1.8), stillbirth (odds ratio [OR] 1.3, 95% CI 1.2–1.5), maternal prolonged hospital stay (OR 4.0, 95% CI 3.4–4.7), maternal death (OR 3.7, 95% CI 2.3–5.9), maternal deaths caused by opioid overdose (OR 5.0, 95% CI 3.4–7.2), and maternal suicide deaths (OR 4.6, 95% CI 3.0–6.6) [10, 11]. However, these studies didn’t separate the effects of prescription opioid use from that of opioid use disorder (OUD) and its treatments. OUD and medication-assisted treatments have been associated with many pregnancy and obstetric complications, such as cesarean delivery (OR 1.2, 95% CI 1.1–1.3), premature rupture of membranes (OR 1.4, 95% CI 1.3–1.6), cardiac arrest (OR 3.6, 95% CI 1.4–9.1), and placental abruption (aOR 2.4, 95% CI 2.1–2.6) [13, 14]. It remains unclear if prescribing opioids to pregnant women is associated with adverse complications during pregnancy or delivery.

Thus, our study aimed to assess the potential for adverse outcomes of prescription opioid use in pregnant women who were enrolled in the Rhode Island (RI) state Medicaid program.

2. Methods

2.1. Data Source

Study data were obtained from the RI Medicaid administrative claims data linked to vital statistics. The data collected by different state agencies were deterministically linked in RI Department of Health (DOH) and RI Executive Office of Health and Human Services using identifiable patient information including patient name and date of birth. The linked data were subsequently deidentified and released for research purposes. About 30% of women in vital statistics were linked to the RI Medicaid claims, which is similar to the proportion of Medicaid-covered population in RI [15]. The Medicaid claims data comprises insurance enrollment; pharmacy dispensing claims consisting of generic name, brand name, date of dispensing, days’ supply, strength, quantity dispensed, and costs; and inpatient and outpatient medical claims containing diagnoses and procedures (recorded by the International Classification of Diseases, 9th and 10th Revisions, Clinical Modification [ICD-9-CM and ICD-10-CM], and Current Procedural Terminology [CPT], Fourth Edition). Data from vital records contained newborns’ birth date, ultrasound-based estimation of gestational age (in complete weeks), birth weights, perinatal conditions, demographic information, and delivery modes. This study was granted exempt status by the Institutional Review Board of University of Rhode Island (IRB 1289357–4) and Rhode Island Department of Health (IRB#: 2019–11).

2.2. Study Cohort

The study cohort consisted of pregnant women aged 12–55 years who had delivered one or multiple live births between 01 January 2008 and 30 September 2015, and who had continuous Medicaid enrollment starting 6 months before pregnancy. Furthermore, we defined loss of Medicaid eligibility during pregnancy as dropout. Estimated beginning of pregnancy was derived using delivery date (from vital statistics) subtracting the gestational age estimated based on ultrasound examinations. Women with OUD, receipt of opioid maintenance treatment (i.e., buprenorphine or methadone), or cancer during baseline (within 6 months preceding pregnancy) or pregnancy (operational definitions of OUD and cancer were presented in eTable 1, see electronic supplementary material [ESM]) were excluded. We excluded cancer patients as cancer may be associated with poorer birth/ pregnancy outcomes and cause confounding by indication. In addition, pregnant women with prescription opioid exposure within 6 months preceding pregnancy were excluded from the study cohort; therefore, the exposure group only includes ‘new users’ of prescription opioids.

2.3. Opioid Exposure

Exposure was defined as having at least one filling of any opioid prescription, which comprised hydrocodone, oxycodone, codeine, tramadol, fentanyl, hydromorphone, meperidine, methadone, morphine, oxymorphone, dihydrocodeine, tapentadol, levorphanol, transdermal buprenorphine, and pentazocine, between the estimated beginning of pregnancy and 1 day before date of delivery or occurrence of outcomes. More specifically, assessment of prescription opioid exposure was started from the first gestational week, and re-assessed and updated in each subsequent gestational week, until delivery. Therefore, patient exposure status may vary in different gestational weeks as individuals defined as exposed in a certain gestational week may switch to unexposed when days’ supply ended, or vice versa. Additionally, we did not evaluate opioid use on the day of delivery to avoid including opioids indicated for pain management during labor.

Exposure to prescription opioids was further characterized in terms of cumulative dose by gestational week and specific opioids. To measure risk of obstetric complications according to the total amount opioid dispensed, we quantified the cumulative dose of prescription opioids in each gestational week and converted this into opioid oral morphine milligram equivalents (MME) [16]. Estimated median of accumulated MME per gestational week was calculated and used as a cutoff point to categorize patients into high and low opioid exposure groups. For each gestational week, we calculated the cumulative MME across different types of prescription opioids. The median of the weekly cumulative MME in the study cohort is 90 mg/week, which was used as the cutoff point for dichotomizing the opioid dispensing dose to low versus high. Effects of specific opioid use on study outcomes have been assessed for most commonly used prescription opioids, such as hydrocodone and oxycodone.

2.4. Outcome Assessment

To comprehensively evaluate perinatal safety of prescription opioid in gestation, we investigated several categories of maternal and obstetric complications occurring during pregnancy to 6 days postpartum: (1) mode of delivery and complications associated with delivery: cesarean delivery, operative vaginal delivery, anesthesia during labor, induction of labor, prolonged birth; (2) placental ischemic complications: placental abruption, premature rupture of membranes, antepartum hemorrhage, and postpartum hemorrhage; (3) acute myocardial infarction or cardiac arrest, thenceforth mentioned as cardiac events; and (4) other common pregnancy complications: antenatal depression and preterm birth. Anesthesia during labor was ascertained from DOH vital records and dichotomized to ‘neuraxial anesthesia’ versus ‘none/local anesthesia’. More specifically, neuraxial anesthesia included spinal or epidural anesthesia. Remaining outcomes were identified using ICD-9 and ICD-10 diagnosis and procedure codes (eTable 2, see ESM) in Medicaid claims data. All individuals were followed until the occurrence of the defined outcome, 6 days after delivery, administrative end of study (October 6, 2015), or dropout defined as loss of enrollment in RI Medicaid, whichever came first.

2.5. Covariates

Time-fixed baseline characteristics were determined at 6-month baseline and included maternal demographic characteristics (age and race), delivery-related factors (i.e., year of birth, multifetal pregnancy), and preexisting comorbid conditions consisting of diabetes mellitus, hypertension, obesity, chronic anemia, tobacco use, alcohol or non-opioid drug abuse or dependence, depression, anxiety/post-traumatic stress disorder (PTSD), and other mental illness, and concomitant medication use that could potentially confound associations between opioid use and maternal or obstetric complications were selected as covariates after a review of the published literature [13, 17–21].

Time-varying confounding factors were ascertained and updated at each gestational week. Two categories of time-varying covariates consisted of (1) the potential indications of opioid prescription (i.e., trauma and pain syndromes) and (2) concomitant medication use that is closely related to opioid use. Trauma and pain syndromes were determined by inpatient/outpatient medical claims on the basis of ICD-9 diagnosis codes and assumed to be present for one gestational week. Concomitant medication use was started on the medication dispensing date and remained as ‘on’ until days’ supply ended. All time-varying covariates were coded in an ‘on’ and ‘off’ manner. To ensure a proper temporal relationship between time-varying covariates and exposure, we lagged the values of time-varying covariates by one time interval (Fig. 1). A complete list of time-invariant and time-varying covariates are presented in Table 1, and their operational definitions are listed in eTable 3 (see ESM).

Fig. 1.

Fig. 1

Open in a new tab

Study design scheme

Table 1.

Baseline characteristic of the study population: RI Medicaid 2008–2015

Characteristics Exposed in pregnancy (n = 545; 5.5%) Unexposed in pregnancy (n = 9278; 94.5%)
Maternal agea [years], mean (SD) 26.9 (5.4) 25.4 (6.1)
Maternal age category, N (%)
 <18 years < 11 (1.3) 699 (7.5)
 18–24 years 198 (36.3) 3748 (40.4)
 25–34 years 289 (53) 4028 (43.4)
 ≥ 35 years 51 (9.4) 803 (8.7)
Race, N (%)
 Black 52 (9.5) 1370 (14.7)
 Hispanic 141 (25.8) 2830 (30.4)
 Other 90 (16.5) 1426 (15.3)
 White 294 (48.3) 3688 (39.6)
Year of birth, N (%)
 2008–2009 99 (18.2) 2428 (26.2)
 2010–2011 132 (24.2) 2593 (27.9)
 2012–2013 165 (30.3) 2154 (23.2)
 2014–2015 149 (27.3) 2103 (22.7)
Multifetal pregnancy, N (%) 22 (4.0) 217 (2.3)
Chronic comorbidities, N (%)
 Diabetes < 11 (1.3) 110 (1.2)
 Obesity 24 (4.4) 400 (4.3)
 Hypertension 17 (3.1) 182 (2)
 Anemia 40 (7.3) 449 (4.8)
 Depression 130 (23.9) 1499 (16.2)
 Anxiety/PTSD 110 (20.1) 1144 (12.3)
 Other mental illnessb 82 (15.0) 864 (9.3)
 Substance/alcohol use disorder 21 (3.9) 270 (2.9)
 Tobacco use disorder 51 (9.4) 502 (5.4)
Concomitant medicationsc (time-varying covariates), N (%)
 Antidiabetic 18 (3.3) 209 (2.3)
 Antihypertensive 18 (3.3) 170 (1.8)
 Antidepressant 84 (15.4) 710 (7.7)
 Antipsychotics 12 (2.2) 99 (1.1)
 Anticonvulsants 22 (4.0) 135 (1.7)
 Anti-anxiety drugs 40 (7.3) 295 (3.2)
Pain indicationsc (time-varying), N (%)
 Pain syndromes overalld 368 (67.5) 4341 (46.8)
 Musculoskeletal pain 143 (26.2) 1125 (12.1)
 Headache/migraine 42 (7.7) 305 (3.3)
 Back pain 179 (32.8) 1863 (20.1)
 Abdominal pain 267 (48.8) 2966 (31.8)
 Fibromyalgia 26 (4.8) 103 (1.1)
 Trauma 88 (16.1) 833 (9)
Open in a new tab

N and % were calculated from time-varying models. Standardized differences and p value for comparisons are not able to be obtained

Small cell count < 11 was suppressed

PTSD post-traumatic stress disorder, RI Rhode Island, SD standard deviation

a

Quadratic form of maternal age at birth as continuous variable was also included in the analysis

b

Included bipolar disorder, mood disorder, schizophrenia, and psychosis

c

Time-varying covariates were evaluated during pregnancy and updated weekly

d

Only a composite of pain syndromes at baseline was included in the analysis

2.6. Statistical Analysis

2.6.1. Marginal Structural Model with Time-varying Exposure and Covariates

2.6.1.1. Motivation

Marginal structural Cox proportional hazards models were fitted to assess the effect of a time-varying exposure to prescription opioids in pregnancy on maternal/pregnancy outcomes with the presence of time-varying covariates. The purpose of applying such models was to address time-varying exposure, covariates, and dropout due to loss of eligibility of Medicaid enrollment. When a time-varying covariate is a predictor for outcome, dropout and subsequent exposure, and is possibly affected by previous exposure, whether or not adjusting for it as a regressor in a standard regression model could lead to a biased effect estimate [22–25]. Marginal structural Cox models have been successfully implemented to assess the effect of various time-varying exposures and tend to yield effect estimates that approximate the findings of randomized control trials [22–25].

2.6.1.2. Time-Varying Exposure and Outcome

To capture the acute effects of opioid use on pregnancy complications and also avoid protopathic bias occurring when opioids are prescribed for treating early symptoms of pregnancy complications, we examined the effects of opioid exposure in two different exposure–outcome time windows in our primary and secondary analyses, respectively [26–28]. In the primary analysis, outcomes were assessed 1 week after the exposure window; as such, exposure was lagged by a 1-week interval (Fig. 1a). In the primary analysis, the follow-up period between opioid use and outcome occurrence is 1–13 days. Considering acute adverse effects of opioids, such as drug-induced cardiac events, may occur within 1–2 days after drug initiation [29], we conducted the secondary analysis in which outcomes were assessed within the same week of opioid exposure, thereby the follow-up period between opioid use and outcome incidence was shortened to 1–6 days (Fig. 1b). In the primary analysis, outcomes that occurred on the delivery date or within 6 days after delivery were assessed for the last exposure window, whereas in the secondary analysis, outcomes occurring in the same week or on the delivery date or within 6 days after delivery were attributed to the last exposure window. To implement marginal structural models (MSMs), analytic data sets were created by transforming the one subject–one record dataset to one with each person-interval (person-week) as a record. Different analytic data sets were created for each outcome of interest.

2.6.1.3. Inverse Probability Treatment Weighting

We fitted two pooled logistic regression models for the numerator and denominator of the treatment model to obtain stabilized inverse probability of treatment weight (IPTW). For the numerator of the stabilized treatment weights, we estimated the probability of prescription opioids received as a function of treatment history (defined as last week’s exposure status), baseline covariates and week of follow-up. For the denominator, we implemented a similar model as for the numerator with the addition of time-varying covariates. For cumulative MME defined as a categorical variable (‘high’, ‘low’, and ‘unexposed’), we used pooled multinomial logistic regression models to estimate treatment weights.

2.6.1.4. Inverse Probability Censoring Weighting

Instead of requiring all women to have continuous enrollment in Medicaid throughout pregnancy, we constructed an inverse probability of censoring weight (IPCW) by means of two pooled logistic regression models, as a way of addressing selection bias due to dropout, defined as loss of eligibility of Medicaid enrollment [22]. Of note, non-censoring was defined as occurrence of events or event-free through to the end of follow-up. The numerator and denominator were the predicted probability of not being censored obtained from two pooled logistic regression models with non-censored as the dependent variable. For the numerator, treatment, baseline covariates, and week of follow-up were adjusted for, whereas treatment, baseline covariates, time-dependent covariates, and week of follow-up were adjusted for the denominator [22, 23].

2.6.2. Final Outcome Model

A product of treatment weights and censoring weights for each person-week in gestation was used as the final weights in the outcome model. We employed weighted pooled logistic models (using generalized equation estimation) for survival, which yielded a consistent estimate of the parameter for the Cox MSM [22]. To address nonlinearity, we included polynomials for continuous variables (i.e., maternal age and week of follow-up) to improve model fit and stability of weights. In the outcome model, exposure, baseline covariates, and weeks of follow-up were adjusted for. The 95% confidence intervals were obtained using a robust sandwich variance estimator. We also examined distribution of the final stabilized inverse probability (IP) weights for each model at each follow-up interval.

2.6.3. Missing Data

For the outcomes (i.e., anesthesia at delivery) that were derived from vital statistics with missing values, we conducted complete case analysis by discarding individuals with missing values in outcome since the fraction of missingness is negligible (about 0.1–0.3%) and the missing mechanism is presumably missing completely at random.

All statistical analyses were conducted using SAS version 9.4 (SAS Institute, Inc., Cary, NC, USA). A two-sided significance level is set as 0.05.

2.7. Sensitivity Analysis

A series of sensitivity analyses were performed to confirm the robustness of our findings given different exposure–outcome time windows, exposure doses, and specific opioid drugs. First of all, to establish a dose–response relationship with regard to maternal opioid use and obstetric outcomes, we re-assessed the association after dichotomizing opioid exposure into ‘low cumulative dose (<median MME/per gestational week)’ and ‘high cumulative dose (≥median MME/per gestational week)’ using the estimated median of cumulative MME among exposed pregnancies as a cutoff point, with unexposed pregnancies as a comparison group. Second, we evaluated the association of obstetric outcomes with exposure to specific opioid generics, categorized into hydrocodone, oxycodone, and codeine, given their high frequencies of being dispensed in pregnancy among Medicaid enrollees. Tramadol use is too rare to be included in the analysis. Dose responses and effects of specific opioids were examined in both primary and secondary analyses.

Lastly, to assess potential confounding by indication, we conducted additional sensitivity analyses that used maternal exposure to opioids in 6-month baseline alone as an active comparator in the sense that women who were dispensed prescription opioids during baseline and discontinued in pregnancy potentially had a similar risk background to women exposed to opioids in pregnancy.

3. Results

A total of 9823 maternal–infant pairs without prescription opioids filled prior to pregnancy were included for analysis. Of these, 9278 (94.5%) mother–infant pairs had no prescription opioid exposure during pregnancy, and 545 (5.5%) mother–infant pairs were exposed to opioid analgesic prescriptions any time throughout pregnancy (Fig. 2). Among 545 exposed pregnancies, 212 (38.9%) received at least one opioid prescription in the first trimester, 199 (36.5%) in the second trimester, and 221 (38.7%) in the third trimester. Sixty-nine (12.6%) received an opioid prescription in any of two trimesters and seven (1.2%) received an opioid prescription during all three trimesters. Among all opioid analgesic prescriptions, hydrocodone (n = 184, 30.3%), oxycodone (n = 141, 23.2%), and codeine (n = 121, 19.9%) were the most commonly prescribed opioids among Medicaid pregnant women between 2010 and 2015. About 90.1% (n = 492) of opioid-exposed pregnant women only received one type of opioid prescription; 8.4% and 1.5% received two and more than two different opioid types, respectively.

Fig. 2.

Fig. 2

Open in a new tab

Flow chart of study cohort

At the time of delivery, the mean age of exposed women was 26.9 years and 25.4 years in the unexposed group. Compared with unexposed, women with exposure to opioids were more frequently white, and had a higher prevalence of pain syndromes, tobacco use, diagnoses of mental disorders, chronic comorbid conditions, and co-medication use for mental disorders during baseline and pregnancy (Table 1). Conversely, the prevalence of diabetes, hypertension, obesity, and substance/alcohol use was similar between the two comparison groups (Table 1).

Table 2 shows the results of primary and secondary analyses using MSMs. In the primary analysis, none of the estimates reached statistical significance, possibly due to wide confidence intervals. In the secondary analysis, the IP-weighted effects of opioid use during pregnancy are statistically significant for antepartum depression (HR 3.19; 95% CI 1.22–8.33), and cardiac events (HR 9.44; 95% CI 1.19–74.83). The incidences of oligohydramnios, preeclampsia, and eclampsia was too low to be analyzed.

Table 2.

Unadjusted and IP-weighted hazard ratios of pregnancy and obstetric outcomes

Outcome Primary analyses
 Secondary analyses
Events occurred 1 week after opioid exposure
 Events occurred within 1 week of opioid exposure
Unadjusted HR (95% CI) IP-weighted HR (95% CI) Unadjusted HR (95% CI) IP-weighted HR (95% CI)
Cesarean delivery, Yes vs No 2.10 (1.32–3.35) 1.70 (0.78–3.67) 1.99 (1.23–3.21) 1.68 (0.74–3.83)
Neuraxial anesthesia during labor vs None/local anesthesiaa 1.53 (1.14–2.07) 1.41 (0.93–2.14) 1.54 (1.14–2.07) 1.40 (0.92–2.14)
Operative vaginal delivery, Yes vs No 1.67 (1.08–2.59) 1.68 (0.81–3477) 1.59 (1.02–2.47) 1.64 (0.76–3.54)
Induction of delivery, Yes vs No 1.17 (0.43–3.14) 0.50 (0.16–1.59) 1.17 (0.44–3.14) 0.56 (0.17–1.79)
Prolonged labor, Yes vs No 1.60 (0.76–3.40) 1.13 (0.41–3.15) 1.61 (0.76–3.41) 1.15 (0.40–3.30)
Placental abruption, Yes vs No 3.65 (0.90–14.80) 2.12 (0.49–26.19) 3.66 (0.90–14.82) 2.07 (0.48–9.05)
Premature rupture of membranes, Yes vs No 1.05 (0.34–3.26) 0.80 (0.24–2.73) 0.70 (0.17–2.81) 0.47 (0.11–2.05)
Antepartum hemorrhage, Yes, vs No 1.23 (0.39–3.84) 0.98 (0.34–2.73) 0.82 (0.20–3.30) 0.61 (0.14–2.64)
Postpartum hemorrhage, Yes vs No 2.01 (0.51–7.96) 4.27 (0.69–26.19) 2.01 (0.51–7.97) 4.19 (0.64–27.35)
Antepartum depression, Yes vs No 1.95 (0.98–3.88) 2.73 (0.96–7.68) 2.20 (1.15–4.20) 3.19 (1.22–8.33)
Cardiac events, Yes vs No 6.10 (0.84–44.39) 3.25 (0.43–24.81) 6.10 (0.84–44.49) 9.44 (1.19–74.83)
Preterm birth, Yes vs No 0.89 (0.29–2.78) 0.55 (0.16–1.92) 0.89 (0.29–2.79) 0.53 (0.15–1.85)
Open in a new tab

Maternal opioid exposure compared with unexposed pregnancies in primary analyses and secondary analyses using time-varying exposure and covariates weighted by propensity score

Adjusted for previous exposure history, baseline covariates (i.e., time interval and polynomial terms, maternal age, quadratic form of maternal age, race, year of birth, multiple gestation, pre-existing comorbid conditions) in the numerator models, and previous exposure history, baseline covariates, time-varying pain indications and comedication use in the denominator models

CI confidence interval, HR hazard ratio, IP inverse probability

a

Anesthesia method at delivery is missing for 23 (0.2%) and only complete cases were analyzed. Neuraxial anesthesia during labor includes spinal or epidural anesthesia

Dose responses were examined for high and low doses of opioid exposure for each gestational week, which were dichotomized by the median of cumulative MME (Table 3). In the primary analysis, high doses of opioids (cumulative MME ≥90 mg) were significantly associated with cesarean delivery (HR 3.97; 95% CI 2.26–6.97), anesthesia during labor (neuraxial vs none/local) (HR 2.12; 95% CI 1.48–3.04), and antepartum depression (HR 6.05; 95% CI 1.79–20.46). Operative vaginal delivery relates to both low- (HR 2.25; 95% CI 1.06–4.79 in MME<90 mg) and high-dose opioid use (HR 2.37; 95% CI 1.32–4.27 in MME ≥90 mg). Postpartum hemorrhage only relates to low-dose opioid use (HR 6.74; 95% CI 1.22–37.15), while its result is not available for high-dose opioid exposure due to a small number of events. In the secondary analysis, significant associations were observed for high-dose opioid use with cesarean delivery (HR 3.97; 95% CI 2.26–6.97), receiving neuraxial anesthesia during labor (HR 2.13; 95% CI 1.46–3.11), operative vaginal delivery (HR 2.75; 95% CI 1.60–4.72), antepartum depression (HR 6.56; 95% CI 2.28–18.93), and cardiac events (HR 27.06; 95% CI 3.13–234.24). Postpartum hemorrhage was only associated with low-dose opioid use (HR 6.78; 95% CI 1.18–38.97), while the estimate is not available due to limited outcome events in the high-dose exposure group (Table 3).

Table 3.

Significant unadjusted and IP-weighted hazard ratios and 95% CI for exposure to low dose (MME < 90 per week) and high dose (≥90 per week) opioids compared with unexposed pregnancies (time-varying exposure and covariates)

Outcome eventsa Low dose (MME < 90/week) vs unexposed
 High dose (MME ≥ 90/week) vs unexposed
Unadjusted HR (95% CI) IP-weighted HR (95% CI) Unadjusted HR (95% CI) IP-weighted HR (95% CI)
Primary analyses: events occurred 1 week after opioid exposure
 Cesarean delivery 1.36 (0.64–2.88) 0.90 (0.33–2.42) 3.22 (1.78–5.83) 3.97 (2.26–6.97)
 Neuraxial anesthesia during labor vs None/local anesthesiab 1.18 (0.77–1.81) 1.48 (0.80–2.77) 2.05 (1.34–3.15) 2.12 (1.48–3.04)
 Operative vaginal delivery, Yes, vs No 1.67 (0.95–2.96) 2.25 (1.06–4.79) 1.67 (0.83–3.34) 2.37 (1.32–4.27)
 Postpartum hemorrhage, Yes, vs No 3.42 (0.85–13.72) 6.74 (1.22–37.15) N/A N/A
 Antepartum depression, Yes, vs No 2.06 (0.86–4.96) 2.08 (0.85–5.10) 1.78 (0.59–5.37) 6.05 (1.79–20.46)
Secondary analyses: events occurred within 1 week of opioid exposure
 Cesarean delivery, Yes, vs No 0.78 (0.29–2.09) 0.56 (0.17–1.82) 3.81 (2.19–6.63) 4.37 (2.59–7.38)
 Neuraxial anesthesia at labor vs None/local anesthesiab 1.18 (0.77–1.81) 1.47 (0.79–2.74) 2.06 (1.34–3.15) 2.13 (1.46–3.11)
 Operative vaginal delivery, Yes, vs No 1.25 (0.66–2.40) 1.72 (0.76–3.88) 2.08 (1.12–3.86) 2.75 (1.60–4.72)
 Postpartum hemorrhage, Yes, vs No 3.42 (0.85–13.72) 6.78 (1.18–38.97) N/A N/A
 Antepartum depression, Yes, vs No 2.07 (0.86–4.97) 2.00 (0.89–4.49) 2.38 (0.92–6.18) 6.56 (2.28–18.93)
 Cardiac events, Yes, vs No N/A N/A 14.79 (2.01–108.54) 27.06 (3.13–234.24)
Open in a new tab

Adjusted for previous exposure history, baseline covariates (i.e., time interval and polynomial terms, maternal age, quadratic form of maternal age, race, year of birth, multiple gestation, pre-existing comorbid conditions) in the numerator models, and previous exposure history, baseline covariates, time-varying pain indications and comedication use in the denominator models

90 MME per gestational week is the estimated median of weekly cumulative MME in the exposed pregnancies

CI confidence interval, HR hazard ratio, IP inverse probability, MME milligram morphine equivalent, N/A the results are not available due to the limited number of outcome events

a

Only present outcomes with a significant association with opioid low or high dose exposure

b

Assuming anesthesia during labor is missing completely at random and only complete cases were analyzed. Neuraxial anesthesia during labor includes spinal or epidural anesthesia

Further subgroup analyses were conducted for specific opioids (Table 4). In the primary analysis, maternal exposure to oxycodone during pregnancy was associated with a higher rate of cesarean delivery (HR 4.05; 95% CI 1.92–8.56), receiving neuraxial anesthesia during labor (HR 1.68; 95% CI 1.31–2.15), operative vaginal delivery (HR 3.53; 95% CI 2.19–5.71); and antepartum depression (HR 13.18; 95% CI 6.03–28.78), while exposure to codeine was associated with a higher risk of receiving neuraxial anesthesia during labor (HR 1.83; 95% CI 1.02–3.29). In the secondary analysis, hydrocodone was related to an increased risk of antepartum depression (HR 3.00; 95% CI 1.02–8.82); and cardiac events (HR 38.73; 95% CI 4.45–336.82). For oxycodone and codeine, similar effects as shown in the primary analysis were observed: oxycodone was significantly associated with cesarean delivery (HR 2.81; 95% CI 1.00–7.88), receiving neuraxial anesthesia during labor (HR 1.62; 95% CI 1.18–2.21), operative vaginal delivery (HR 4.15; 95% CI 2.32–7.43), and antepartum depression (HR 11.93; 95% CI 5.50–25.87), and codeine had significant effect on neuraxial anesthesia during labor (HR 1.81; 95% CI 1.10–2.98).

Table 4.

Significant unadjusted and IP-weighted hazard ratios and 95% CI for specific opioids compared with unexposed pregnancies (time-varying exposure and covariates)

Outcome Primary analyses
 Secondary analyses
Events occurred 1 week after opioid exposure
 Events occurred within 1 week of opioid exposure
Unadjusted HR (95% CI) IP-Weighted HR (95% CI) Unadjusted HR (95% CI) IP-Weighted HR (95% CI)
Cesarean delivery, Yes vs No
 Hydrocodone vs unexposed 2.35 (1.01–5.40) 2.15 (0.52–8.57) 2.30 (1.09–4.85) 2.69 (0.88–8.24)
 Oxycodone vs unexposed 2.12 (0.83–5.39) 4.05 (1.92–8.56) 1.93 (0.83–4.53) 2.81 (1.00–7.88)
 Codeine vs unexposed 1.62 (0.66–3.94) 1.24 (0.48–3.29) 1.94 (0.92–4.10) 1.43 (0.61–3.33)
Neuraxial anesthesia during labor vs None/local anesthesiab
 Hydrocodone vs unexposed 1.06 (0.54–2.08) 1.84 (0.72–4.70) 1.16 (0.65–2.06) 1.51 (0.81–2.84)
 Oxycodone vs unexposed 1.53 (0.83–2.82) 1.68 (1.31–2.15) 1.32 (0.74–2.34) 1.62 (1.18–2.21)
 Codeine vs unexposed 1.88 (1.21–2.93) 1.83 (1.02–3.29) 1.96 (1.31–2.93) 1.81 (1.10–2.98)
Operative vaginal delivery, Yes, vs No
 Hydrocodone vs unexposed 1.18 (0.44–3.17) 1.47 (0.52–4.23) 0.74 (0.24–2.34) 0.65 (0.21–2.01)
 Oxycodone vs unexposed 2.58 (1.25–5.34) 3.53 (2.19–5.71) 1.97 (0.97–3.99) 4.15 (2.32–7.43)
 Codeine vs unexposed 1.09 (0.46–2.62) 0.81 (0.28–2.35) 1.52 (0.78–2.96) 1.43 (0.55–3.70)
Antepartum depression, yes, vs no
 Hydrocodone vs unexposed 1.61 (0.40–6.48) 1.87 (0.50–6.97) 2.02 (0.65–6.21) 3.00 (1.02–8.82)
 Oxycodone vs unexposed 4.36 (1.74–10.96) 13.18 (6.03–28.78) 4.08 (1.76–9.42) 11.93 (5.50–25.87)
 Codeine vs unexposed 1.36 (0.34–5.46) 1.21 (0.31–4.73) 1.18 (0.29–4.77) 0.86 (0.22–3.43)
Cardiac events, Yes, vs No
 Hydrocodone vs unexposed N/A N/A 18.13 (2.46–133.55) 38.73 (4.45–336.82)
 Oxycodone vs unexposed N/A N/A N/A N/A
 Codeine vs unexposed N/A N/A N/A N/A
Open in a new tab

Adjusted for previous exposure history, baseline covariates (i.e., time interval and polynomial terms, maternal age, quadratic form of maternal age, race, year of birth, multiple gestation, pre-existing comorbid conditions) in the numerator models, and previous exposure history, baseline covariates, time-varying pain indications and comedication use in the denominator models

90 MME per gestational week is the estimated median of weekly cumulative MME in the exposed pregnancies

CI confidence interval, HR hazard ratio, IP inverse probability, N/A the results are not available due to the limited number of outcome events

a

Only present outcomes with a significant association with opioid low or high dose exposure

b

Assuming anesthesia during labor is missing completely at random and only complete cases were analyzed. Neuraxial anesthesia during labor includes spinal or epidural anesthesia

Lastly, we re-assessed the effects of maternal prescription opioids use using an active comparator, women prescribed opioids at baseline and discontinued in pregnancy. Although none of the estimates are statistically significant due to a smaller sample size in the comparison group, the magnitudes of the estimates are similar and in the same direction as the results from the original analysis (eTable 4, see ESM). The dose–response relationship was maintained largely consistent with results that used non-exposed pregnancies as reference in both primary and secondary analyses, while the magnitude of associations were slightly diminished (eTable 5).

4. Discussion

In our analysis of RI Medicaid-enrolled opioid-naïve pregnant women, about 5.5% women were dispensed prescription opioids during pregnancy. Significant associations were observed between opioid exposure during pregnancy and increased risks of cesarean delivery, operative vaginal delivery, antepartum depression, and cardiac events that occurred during the week of opioid exposure. Furthermore, our dose–response analysis suggested an increased risk of cesarean delivery, neuraxial anesthesia during labor, operative vaginal delivery, antepartum depression, and cardiac events in pregnant women with high-dose opioid use within a couple of weeks. Oxycodone use was associated with most of these adverse obstetric outcomes, except for cardiac events, which showed significantly higher risk with hydrocodone use in pregnant women.

Our findings expand on the prior available evidence. Our study further confirmed the significant association of maternal prescription opioid use on risks of cesarean section and postpartum hemorrhage (dosage < 90 MME/ week) among opioid-naive women; this distinction is important since women with opioid dependence or abuse or those using maintenance prescription opioids have been shown to have different risks of adverse peripartum outcomes [30]. It is not surprising that we identified an increased use of neuraxial anesthesia, including spinal or epidural anesthesia, in women with high-dose opioid exposure compared with unexposed women, since neuraxial anesthesia was suggested for peripartum pain management in opioid-tolerant parturients [31]. In addition, we identified an increased hazard of operative vaginal delivery in women exposed to opioids during pregnancy compared with non-users. Usually parturients have operative vaginal deliveries due to fetal distress, which relates to maternal opioid abuse [32, 33].

Depression, pain severity, and opioid use are intertwined and affect each other profoundly [34–39]. Use of higher opioid dose was associated with an increased risk of depression [40–42]. In our study, effect of opioid use during pregnancy on antepartum depression was apparent in the secondary analysis, and the risk of antepartum depression increased with higher dose of opioid use (≥ 90 MME/day) after adjusting for history of depression and psychotropic medication use in pregnancy. Although previous studies have reported that opioid abuse or addiction is a risk factor for placental abruption [43–48], similar results were not observed in our study. This suggests that prolonged exposure to opioids may be required for adverse peripartum outcomes related to the placenta.

Opioid-related acute cardiac events have also been observed in published data [13, 49]. One of the possible explanations is that opioid administration combined with other medications (e.g., benzodiazepine) may potentially lead to hypotension and compromised respiratory function, thus resulting in decreased cardiac function [50, 51]. Additionally, studies have suggested a number of non-maintenance prescription opioids (e.g., tramadol and oxycodone) and maintenance prescription opioids (e.g., methadone) are intermediate- to high-risk drugs that may potentially induce prolongation of the QT intervals, potentially causing fatal arrhythmias [50, 52–56]. Our analysis showed a significant association between opioid use and cardiac events occurred within 1 week of opioid dispensing, with higher dose exposure, and oxycodone or hydrocodone use. Our findings are rational as demonstrated by previous studies that drug-related cardiac events are an acute event that occurred soon following drug use [57–59].

In our data, effects of opioid use during pregnancy on preterm birth were not statistically significant. The null associations observed might be partially attributed to the rare events and a relatively small study cohort as prior studies have confirmed an increased risk of preterm delivery in pregnant women exposed to prescription opioids compared with their counterparts [13, 60, 61]. We observed a seemingly protective effect of opioid use in pregnancy in preterm birth in the secondary analyses, while not statistically significant. This protective association could have been obtained by chance or due to protopathic bias, rare events, or short follow-up time (1–6 days).

The fact that there is a clear association between iatrogenic opioid exposure in pregnancy and perinatal maternal complications adds urgency to national initiatives aimed towards reducing prescription opioid use during pregnancy. ACOG recommends that opioid analgesia be reserved only for women in severe pain [62], but this and other organizations must disseminate their practice guidelines to educate all providers who care for women in pregnancy—including in an emergency setting—that non-opioid modalities for pain management should be first-line during pregnancy. Only in the setting of a failed first-line non-narcotic regimen and persistent pain does the benefit of maternal opioid exposure clearly outweigh the risk of maternal perinatal complications.

4.1. Limitations

Several study limitations exist. First, association, instead of causation, was estimated in our analysis. Even though we addressed the most relevant risk factors [22, 62–64], information on confounders such as pre-pregnancy BMI, previous pregnancy outcomes (e.g., cesarean section, preterm birth, miscarriage), and over-the-counter drug use were poorly collected and thus not adjusted for in our analyses. Additionally, socioeconomic status (e.g., education/income) was not accounted for, although Medicaid enrollees tend to represent a subpopulation with low-socioeconomic status. Second, confounding by indication may still exist since acute or chronic pain in the head, abdomen, pelvis, or lower back can be a warning sign of underlying medical conditions that lead to inferior pregnancy and obstetric complications [65–67]. The sensitivity analysis using an active comparator as reference might suggest the presence of confounding by indication in the current study, however, such unmeasured confounders are less likely to thoroughly explain the observed risk of several adverse pregnancy outcomes associated with high cumulative opioid exposure. Third, this study is subject to a limited sample size, particularly for certain rare outcomes. As a result, chance findings cannot be ruled out and the current study should be replicated with a larger scaled and more diverse population, although results on sensitivity analyses were reasonably consistent. The event frequencies in exposed and unexposed women are not listed in the tables because for most of the events frequencies were < 11, and therefore could not be presented as per the data agreement with the RI DOH. Fourth, pharmacy claims dispensing does not necessarily indicate use of opioids in a realworld setting. Actual use of opioids may differ from the claims data. However, the resulting exposure misclassification is likely non-differential. Fifth, since there are only 1–13 or 1–6 days from opioid exposure to outcome occurrence in this study, the estimated effects are more plausible for acute events, such as cardiac events, that have a short latent period. Lastly, our study cohort is based on RI Medicaid enrollees, including a total of 9823 mother–infant dyads, which is about 11.4% of RI Medicaid-covered live births that were born between 01 January 2008 and 30 September 2015 in mothers aged 12–55 years (Fig. 2). The generalizability of the findings may be limited to US women in the New England area with lower socioeconomic status.

5. Conclusions

In our findings, high cumulative opioid use in pregnant women was associated with an increased risk of multiple pregnancy and obstetric complications, such as cesarean delivery, operative vaginal delivery, neuraxial anesthesia during labor, antepartum depression, and cardiac events. A cautious approach is needed when prescribing cumulatively high doses of opioids to pregnant women.

Supplementary Material

Supplementary files

Key Points.

Prescribing opioids to pregnant women is associated with an increased risk of pregnancy and obstetrical complications.

A cautious approach is needed when prescribing opioids, especially high doses, to pregnant women.

Applying Marginal Structural Models in pregnancy studies helps address time-dependent exposure and covariates.

Acknowledgements

We thank the Rhode Island Department of Health and the Executive Office of Health and Human Services for providing the data access. Samara VinerBrown, MS, and William Arias, MPH, provided comments on the clarification of study population. Samara VinerBrown, MS, William Arias, MPH, and Ellen Amore, MS in the Center for Health Data & Analysis, Rhode Island Department of Health, and Dr Rebecca Lebeau’s team in Rhode Island Executive Office of Human and Health Services provided data linkage, data cleaning, and data preparation. No financial compensation was provided.

Funding

The study is funded by grants from the National Institute of Health (NIH)/National Institute of Child Health and Human Development (NIH/NICHD Grant number: 1R15HD097588-01; Principal Investigator: Xuerong Wen). The funding source had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Conflicts of interest/Competing interests

Dr Meador has received research support from the National Institutes of Health and Sunovion Pharmaceuticals, and travel support from UCB Pharma. The Epilepsy Study Consortium pays Dr Meador’s university for his research consultant time related to Eisai, GW Pharmaceuticals, NeuroPace, Novartis, Supernus, Upsher-Smith Laboratories, and UCB Pharma. Other authors, XW, SW, AKL, KEW, and ECB have no conflicts of interests to declare.

Footnotes

Availability of data and material The study data was linked and deidentified by the RI DOH and RI EOHHS. The deidentified data was released to Dr Wen’s research team as per the Data Utilization Agreement.

Code availability Not applicable.

Ethics approval This study was considered exempt by the RI DOH IRB and URI IRB.

Consent to participate Not applicable.

Consent for publication Not applicable.

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s40264-021-01115-6.

References

  • 1.Opioid use and opioid use disorder in pregnancy|ACOG. 2020. https://www.acog.org/clinical/clinical-guidance/committee-opinion/articles/2017/08/opioid-use-and-opioid-use-disorder-in-pregnancy. Accessed 22 Jun 2020.
  • 2.Wen X, Belviso N, Lebeau R, et al. Prescription opioid use among pregnant women enrolled in Rhode Island Medicaid. R I Med J (2013) 2019;102(6):35–40. [PMC free article] [PubMed] [Google Scholar]
  • 3.Bateman BT, Hernandez-Diaz S, Rathmell JP, et al. Patterns of opioid utilization in pregnancy in a large cohort of commercial insurance beneficiaries in the United States. Anesthesiology. 2014;120(5):1216–24. 10.1097/ALN.0000000000000172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Desai RJ, Hernandez-Diaz S, Bateman BT, Huybrechts KF. Increase in prescription opioid use during pregnancy among medicaid-enrolled women. Obstet Gynecol. 2014;123(5):997–1002. 10.1097/AOG.0000000000000208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Epstein RA, Bobo WV, Martin PR, et al. Increasing pregnancy-related use of prescribed opioid analgesics. Ann Epidemiol. 2013;23(8):498–503. 10.1016/j.annepidem.2013.05.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Broussard CS, Rasmussen SA, Reefhuis J, et al. Maternal treatment with opioid analgesics and risk for birth defects. Am JObstet Gynecol 2011;204(4):314.e1–11. 10.1016/j.ajog.2010.12.039. [DOI] [PubMed] [Google Scholar]
  • 7.Yazdy MM, Mitchell AA, Tinker SC, Parker SE, Werler MM. Periconceptional use of opioids and the risk of neural tube defects. Obstet Gynecol 2013;122(4):838–44. 10.1097/AOG.0b013e3182a6643c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Fishman B, Daniel S, Koren G, Lunenfeld E, Levy A. Pregnancy outcome following opioid exposure: a cohort study. PLoS ONE. 2019;14(7): e0219061. 10.1371/journal.pone.0219061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Delate T, Hodges M, Swank M, Ota T, Pratt CL. Birth outcomes with prescribed chronic and acute opioid exposure during pregnancy. J Opioid Manag 2019;15(5):407–15. 10.5055/jom.2019.0529. [DOI] [PubMed] [Google Scholar]
  • 10.Whiteman VE, Salemi JL, Mogos MF, Cain MA, Aliyu MH, Salihu HM. Maternal opioid drug use during pregnancy and its impact on perinatal morbidity, mortality, and the costs of medical care in the United States. J Pregnancy 2014. 10.1155/2014/906723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Metz TD, Rovner P, Hoffman MC, Allshouse AA, Beckwith KM, Binswanger IA. Maternal deaths from suicide and overdose in Colorado, 2004–2012. Obstet Gynecol 2016;128(6):1233–40. 10.1097/AOG.0000000000001695. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Rosoff DB, Smith GD, Lohoff FW. Prescription opioid use and risk for major depressive disorder and anxiety and stress-related disorders. JAMA Psychiat. 2021;78(2):1–10. 10.1001/jamapsychiatry.2020.3554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Maeda A, Bateman BT, Clancy CR, Creanga AA, Leffert LR. Opioid abuse and dependence during pregnancy: temporal trends and obstetrical outcomes. Anesthesiology. 2014;121(6):1158–65. 10.1097/ALN.0000000000000472. [DOI] [PubMed] [Google Scholar]
  • 14.Kaltenbach K, Berghella V, Finnegan L. Opioid dependence during pregnancy: effects and management. Obstet Gynecol Clin North Am. 1998;25(1):139–51. 10.1016/S0889-8545(05)70362-4. [DOI] [PubMed] [Google Scholar]
  • 15.Medicaid in Rhode Island. 2020; p. 2. https://files.kff.org/attachment/fact-sheet-medicaid-state-RI. Accessed 17 June 2021. [Google Scholar]
  • 16.Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain—United States, 2016. JAMA. 2016;315(15):1624–1645. 10.1001/jama.2016.1464 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Nezvalová-Henriksen K, Spigset O, Nordeng H. Effects of codeine on pregnancy outcome: results from a large population-based cohort study. Eur J Clin Pharmacol. 2011;67(12):1253–61. 10.1007/s00228-011-1069-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Ananth CV, Peltier MR, Kinzler WL, Smulian JC, Vintzileos AM. Chronic hypertension and risk of placental abruption: is the association modified by ischemic placental disease? Am J Obstet Gynecol 2007;197(3):273.e1–273.e7. 10.1016/j.ajog.2007.05.047. [DOI] [PubMed] [Google Scholar]
  • 19.Yanit KE, Snowden JM, Cheng YW, Caughey AB. The impact of chronic hypertension and pregestational diabetes on pregnancy outcomes. Am J Obstet Gynecol. 2012;207(4):333.e1–6. 10.1016/j.ajog.2012.06.066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Borthen I Obstetrical complications in women with epilepsy. Seizure. 2015;28:32–4. 10.1016/j.seizure.2015.02.018. [DOI] [PubMed] [Google Scholar]
  • 21.Dubovicky M, Belovicova K, Csatlosova K, Bogi E. Risks of using SSRI/SNRI antidepressants during pregnancy and lactation. Interdiscip Toxicol 2017;10(1):30–4. 10.1515/intox-2017-0004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology 2000;11(5):561–70. 10.1097/00001648-200009000-00012. [DOI] [PubMed] [Google Scholar]
  • 23.Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate the joint causal effect of nonrandomized treatments. J Am Stat Assoc. 2001;96(454):440–448. 10.1198/016214501753168154 [DOI] [Google Scholar]
  • 24.Choi HK, Hernán MA, Seeger JD, Robins JM, Wolfe F. Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study. Lancet. 2002;359(9313):1173–7. 10.1016/S0140-6736(02)08213-2. [DOI] [PubMed] [Google Scholar]
  • 25.Cook NR, Cole SR, Hennekens CH. Use of a marginal structural model to determine the effect of aspirin on cardiovascular mortality in the Physicians’ Health Study. Am J Epidemiol. 2002;155(11):1045–53. 10.1093/aje/155.11.1045. [DOI] [PubMed] [Google Scholar]
  • 26.Faillie J-L. Indication bias or protopathic bias? Br J Clin Pharmacol 2015;80(4):779–80. 10.1111/bcp.12705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Patrick SW, Dudley J, Martin PR, et al. Prescription opioid epidemic and infant outcomes. Pediatrics. 2015. 10.1542/peds.2014-3299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Tamim H, Monfared AAT, LeLorier J. Application of lag-time into exposure definitions to control for protopathic bias. Pharmacoepidemiol Drug Saf 2007;16(3):250–8. 10.1002/pds.1360. [DOI] [PubMed] [Google Scholar]
  • 29.Sachinidis A Cardiotoxicity and heart failure: lessons from human-induced pluripotent stem cell-derived cardiomyocytes and anticancer drugs. Cells. 2020;9:4. 10.3390/cells9041001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.The ASAM. National practice guideline for the treatment of opioid use disorder: 2020 focused update. J Addict Med 2020;14(2S):1–91. 10.1097/ADM.0000000000000633. [DOI] [PubMed] [Google Scholar]
  • 31.Raymond BL, Kook BT, Richardson MG. The opioid epidemic and pregnancy: implications for anesthetic care. Curr Opin Anesthesiol 2018;31(3):243–50. 10.1097/ACO.0000000000000590. [DOI] [PubMed] [Google Scholar]
  • 32.Keikha F, Vahdani FG, Latifi S. The effects of maternal opium abuse on fetal heart rate using non-stress test. Iran J Med Sci 2016;41(6):479–85. [PMC free article] [PubMed] [Google Scholar]
  • 33.Bell J, Towers CV, Hennessy MD, Heitzman C, Smith B, Chattin K. Detoxification from opiate drugs during pregnancy. Am J Obstet Gynecol 2016;215(3):374.e1–374.e6. 10.1016/j.ajog.2016.03.015. [DOI] [PubMed] [Google Scholar]
  • 34.Hooten WM, Shi Y, Gazelka HM, Warner DO. The effects of depression and smoking on pain severity and opioid use in patients with chronic pain. Pain. 2011;152(1):223–9. 10.1016/j.pain.2010.10.045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Bär K-J, Brehm S, Boettger MK, Boettger S, Wagner G, Sauer H. Pain perception in major depression depends on pain modality. Pain. 2005;117(1–2):97–103. 10.1016/j.pain.2005.05.016. [DOI] [PubMed] [Google Scholar]
  • 36.de Souza JB, Potvin S, Goffaux P, Charest J, Marchand S. The deficit of pain inhibition in fibromyalgia is more pronounced in patients with comorbid depressive symptoms. Clin J Pain. 2009;25(2):123–7. 10.1097/AJP.0b013e318183cfa4. [DOI] [PubMed] [Google Scholar]
  • 37.Wasan AD, Davar G, Jamison R. The association between negative affect and opioid analgesia in patients with discogenic low back pain. Pain. 2005;117(3):450–61. 10.1016/j.pain.2005.08.006. [DOI] [PubMed] [Google Scholar]
  • 38.Wasan AD, Kaptchuk TJ, Davar G, Jamison RN. The association between psychopathology and placebo analgesia in patients with discogenic low back pain. Pain Med. 2006;7(3):217–28. 10.1111/j.1526-4637.2006.00154.x. [DOI] [PubMed] [Google Scholar]
  • 39.Currie SR, Wang J. More data on major depression as an antecedent risk factor for first onset of chronic back pain. Psychol Med 2005;35(9):1275–82. 10.1017/S0033291705004952. [DOI] [PubMed] [Google Scholar]
  • 40.Scherrer JF, Salas J, Lustman PJ, Burge S, Schneider FD. Change in opioid dose and change in depression in a longitudinal primary care patient cohort. Pain. 2015;156(2):348–55. 10.1097/01.j.pain.0000460316.58110.a0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Scherrer JF, Salas J, Copeland LA, et al. Prescription opioid duration, dose, and increased risk of depression in 3 large patient populations. Ann Fam Med 2016;14(1):54–62. 10.1370/afm.1885. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.King S Depression, opioid analgesics, and pregnancy: what we know and what we don’t. Patient care online. 2020. https://www.patientcareonline.com/view/depression-opioid-analgesics-and-pregnancy-what-we-know-and-what-we-dont. Accessed 15 Jul 2020.
  • 43.Ghaheh HS, Feizi A, Mousavi M, Sohrabi D, Mesghari L, Hosseini Z. Risk factors of placental abruption. J Res Med Sci 2013;18(5):422–6. [PMC free article] [PubMed] [Google Scholar]
  • 44.Maeda A, Creanga AA. Opioid abuse and dependence during pregnancy. Perioperative Med. 2020;2020:8. [Google Scholar]
  • 45.Salari Z, Mirzaie F, Mehran M. Evaluation of relationship between opioid addiction and placental abruption. Avicenna J Clin Med. 2007;14(1):39–43. [Google Scholar]
  • 46.Bhuvaneswar CG, Chang G, Epstein LA, Stern TA. Cocaine and opioid use during pregnancy: prevalence and management. Prim Care Companion J Clin Psychiatry. 2008;10(1):59–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Minnes S, Lang A, Singer L. Prenatal tobacco, marijuana, stimulant, and opiate exposure: outcomes and practice implications. Addict Sci Clin Pract. 2011;6(1):57–70. [PMC free article] [PubMed] [Google Scholar]
  • 48.Goler NC, Armstrong MA, Taillac CJ, Osejo VM. Substance abuse treatment linked with prenatal visits improves perinatal outcomes: a new standard. J Perinatol. 2008;28(9):597–603. 10.1038/jp.2008.70. [DOI] [PubMed] [Google Scholar]
  • 49.Salihu HM, Salemi JL, Aggarwal A, et al. Opioid drug use and acute cardiac events among pregnant women in the United States. Am J Med. 2018;131(1):64–71.e1. 10.1016/j.amjmed.2017.07.023. [DOI] [PubMed] [Google Scholar]
  • 50.Chen A, Ashburn MA. Cardiac effects of opioid therapy. Pain Med. 2015;16(suppl_1):S27–31. 10.1111/pme.12915. [DOI] [PubMed] [Google Scholar]
  • 51.Pattinson KTS. Opioids and the control of respiration. Br J Anaesth. 2008;100(6):747–58. 10.1093/bja/aen094. [DOI] [PubMed] [Google Scholar]
  • 52.Sudhakaran S, Surani SS, Surani SR. Prolonged ventricular asystole: a rare adverse effect of hydrocodone use. Am J Case Rep 2014;15:450–3. 10.12659/AJCR.891214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Fareed A, Patil D, Scheinberg K, et al. Comparison of QTc interval prolongation for patients in methadone versus buprenorphine maintenance treatment: a 5-year follow-up. J Addict Dis 2013;32(3):244–51. 10.1080/10550887.2013.824333. [DOI] [PubMed] [Google Scholar]
  • 54.Bogen DL, Hanusa BH, Perel JM, Sherman F, Mendelson MA, Wisner KL. Corrected QT interval and methadone dose and concentrations in pregnant and postpartum women. J Clin Psychiatry 2017;78(8):e1013–9. 10.4088/JCP.16m11318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Behzadi M, Joukar S, Beik A. Opioids and cardiac arrhythmia: a literature review. MPP 2018;27(5):401–14. 10.1159/000492616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Marteau D, McDonald R, Patel K. The relative risk of fatal poisoning by methadone or buprenorphine within the wider population of England and Wales. BMJ Open. 2015;5(5): e007629. 10.1136/bmjopen-2015-007629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Mittleman-Murray A, Lewis-Rebecca A, Maclure M, Sherwood-Jane B, Muller-James E. Triggering myocardial infarction by marijuana. Circulation. 2001;103(23):2805–9. 10.1161/01.CIR.103.23.2805. [DOI] [PubMed] [Google Scholar]
  • 58.Vozoris NT, Wang X, Austin PC, et al. Adverse cardiac events associated with incident opioid drug use among older adults with COPD. Eur J Clin Pharmacol 2017;73(10):1287–95. 10.1007/s00228-017-2278-3. [DOI] [PubMed] [Google Scholar]
  • 59.Frid AA, Matthews EJ. Prediction of drug-related cardiac adverse effects in humans-B: use of QSAR programs for early detection of drug-induced cardiac toxicities. Regul Toxicol Pharmacol 2010;56(3):276–89. 10.1016/j.yrtph.2009.11.005. [DOI] [PubMed] [Google Scholar]
  • 60.Sujan AC, Quinn PD, Rickert ME, et al. Maternal prescribed opioid analgesic use during pregnancy and associations with adverse birth outcomes: a population-based study. PLoS Med 2019;16(12): e1002980. 10.1371/journal.pmed.1002980. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Azuine RE, Ji Y, Chang H-Y, et al. Prenatal risk factors and perinatal and postnatal outcomes associated with maternal opioid exposure in an urban, low-income, multiethnic US population. JAMA Netw Open. 2019;2(6): e196405. 10.1001/jamanetworkopen.2019.6405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Robins JM, Hernán MÁ, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology 2000;11(5):550–60. 10.1097/00001648-200009000-00011. [DOI] [PubMed] [Google Scholar]
  • 63.Cole SR, Frangakis CE. The consistency statement in causal inference: a definition or an assumption? Epidemiology 2009;20(1):3–5. 10.1097/EDE.0b013e31818ef366. [DOI] [PubMed] [Google Scholar]
  • 64.Westreich D, Cole SR. Invited commentary: positivity in practice. Am J Epidemiol 2010;171(6):674–677. 10.1093/aje/kwp436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.Ray-Griffith SL, Wendel MP, Stowe ZN, Magann EF. Chronic pain during pregnancy: a review of the literature. Int J Womens Health. 2018;10:153–64. 10.2147/IJWH.S151845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Warning Signs During Pregnancy. Stanford Children’s Health. 2020. https://www.stanfordchildrens.org/en/topic/default?id=warning-signs-during-pregnancy-85-P01199. Accessed 24 July 2020. [Google Scholar]
  • 67.Back Pain During Pregnancy. 2020. https://www.acog.org/en/PatientResources/FAQs/Pregnancy/BackPainDuringPregnancy. Accessed 24 Jul 2020.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary files
NIHMS1771419-supplement-Supplementary_files.docx (34.7KB, docx)

RESOURCES