Drug and Alcohol Dependence 149 (2015) 225–231

Contents lists available at ScienceDirect

Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep

Medication assisted treatment discontinuation in pregnant and postpartum women with opioid use disorder Christine Wilder a,b,∗ , Daniel Lewis a , Theresa Winhusen a a Addiction Sciences Division, Department of Psychiatry and Behavioral Neuroscience, University of Cincinnati College of Medicine, 3131 Harvey Avenue, Cincinnati, OH 45229, USA b Department of Veterans Affairs Medical Center, 3200 Vine Street, Cincinnati, OH 45220, USA

a r t i c l e

i n f o

Article history: Received 14 August 2014 Received in revised form 4 February 2015 Accepted 6 February 2015 Available online 19 February 2015 Keywords: Opioid use disorder Pregnancy Postpartum Medication assisted treatment Methadone Patient dropout

a b s t r a c t Background: Increasing use of opioids has led to an increase in the number of pregnant and postpartum women in medication assisted treatment (MAT) for opioid use disorder. Methods: We (1) conducted a systematic review of published literature on MAT discontinuation (methadone and buprenorphine) in pregnant and postpartum women and (2) determined methadone discontinuation rates in a retrospective cohort (2006–2013) of pregnant and postpartum women in a university affiliated methadone clinic. Results: We found limited generalizable literature reports of discontinuation rates, with a range of prenatal discontinuation rates from 0 to 33% and rates which spanned various prenatal and postnatal periods from 26 to 64%. In our cohort of 229 women, 251 pregnancies were reported, with a prenatal methadone discontinuation rate of 11.0%. Based on a Cox proportional hazards model controlling for age, pregnancy outcome, and duration of treatment prior to delivery, the probability of methadone discontinuation at or before 6 months postpartum was 56.0%. Duration of methadone treatment prior to delivery was inversely associated with risk for postpartum discontinuation of treatment (HR = 0.98, 95% CI (0.96, 0.99)). Conclusions: We conclude that the postpartum period is a time of increased risk for discontinuation of MAT. More accurate assessment of rates of pre- and postpartum MAT discontinuation, as well as further investigation of factors affecting these rates, is warranted. Development and testing of interventions to encourage early prenatal enrollment in MAT and improve postnatal retention in MAT would benefit pregnant women and new mothers with opioid use disorder. Published by Elsevier Ireland Ltd.

1. Introduction The percentage of pregnant women who use opioids has tripled in the last ten years, with 1.2% of all pregnant women reporting opioid use in 2012 (Substance Abuse and Mental Health Services Administration, 2003, 2013). Pregnant women face severe, longterm consequences of drug use, including an increased likelihood of assault and abuse, contracting HIV or hepatitis, miscarriage, delivering infants with physical and behavioral impairments, postpartum depression, and loss of custody of their children (Dakof et al., 2003; Dansky et al., 1999; Holbrook and Kaltenbach, 2012; Kissin et al., 2001). Pregnant women using illicit opioids expose themselves and their fetuses to additional risks specific to

∗ Corresponding author at: Addiction Sciences Division, Department of Psychiatry and Behavioral Neuroscience, University of Cincinnati College of Medicine, 3131 Harvey Avenue, Cincinnati, OH 45229, USA. Tel.: +1 513 585 8285; fax: +1 513 585 8278. E-mail address: [email protected] (C. Wilder). http://dx.doi.org/10.1016/j.drugalcdep.2015.02.012 0376-8716/Published by Elsevier Ireland Ltd.

opioid use, including pregnancy complications, premature labor, low birth weight, and spontaneous abortion (American Congress of Obstetricians and Gynecologists, 2012; Hulse et al., 1998; Kandall et al., 1999; Ludlow et al., 2004). Medication assisted treatment (MAT) with methadone or, more recently, buprenorphine is the most effective treatment for opioid use disorder (OUD) in pregnant women (American Congress of Obstetricians and Gynecologists, 2012; Jones et al., 2012; Minozzi et al., 2013) and those in MAT have better maternal and fetal outcomes than those not enrolled in treatment (Burns et al., 2007; Fajemirokun-Odudeyi et al., 2006; Welle-Strand et al., 2013). Some studies have also shown a relationship between length of time in methadone treatment and improved fetal outcomes (Burns et al., 2007; Peles et al., 2012). Although the effectiveness of MAT exclusively in postpartum women has not been studied, MAT is considered a highly effective treatment for OUD in the general non-pregnant population (Mattick et al., 2009), which includes postpartum women. Because MAT has such strong evidence for improved treatment outcomes, retention of patients in MAT is critical. Treatment

226

C. Wilder et al. / Drug and Alcohol Dependence 149 (2015) 225–231

retention in MAT has been studied extensively in the non-pregnant population. Women (Kelly et al., 2011; Newman et al., 1976), older patients (Deck and Carlson, 2005; MacGowan et al., 1996; Magura et al., 1998; Newman et al., 1976), individuals who do not use other drugs during treatment (MacGowan et al., 1996; Magura et al., 1998), and individuals with no criminal justice involvement (Cox et al., 2013; Deck and Carlson, 2005; Greenfield et al., 1996; Magura et al., 1998; McGuire, 1979) are more likely to remain in methadone treatment for 90 days or more. Additionally, treatment with higher doses of methadone (≥60 mg per day) has been associated consistently with better treatment retention (Bao et al., 2009; Brady et al., 2005; Faggiano et al., 2003). Surprisingly, however, there has been limited focus on MAT retention in pregnant or postpartum women, despite the fact that substance-using pregnant women can be difficult to retain in treatment (Haller et al., 1997; Jones et al., 2002). Studies of interventions to improve MAT retention in pregnant and postpartum women are also limited. In this article, we first review the published literature on MAT discontinuation rates in pregnant and postpartum women and on interventions to improve treatment retention in this group. We then present new data on treatment discontinuation from a methadone program treating pregnant and postpartum women. Finally, we synthesize these findings with previously published results and make recommendations for future research.

centers of several of the city’s main hospitals and therefore receives referrals for opioid dependent pregnant women wishing to enter methadone treatment. The clinic collaborates directly with prenatal treatment providers but no prenatal treatment occurs on site. Pregnant women receive treatment which includes methadone dosing, weekly individual counseling, biweekly physician appointments with dose adjustments, biweekly urine drug testing, and coordination of care with their prenatal care providers. Pregnant women are eligible for significantly reduced fees if they fulfill treatment requirements and provide drug-free urine tests, although these fee reductions continue only for the duration of the pregnancy and are not available postpartum. After hospital discharge following delivery, women are assessed every three days by the clinic physician for dose adjustment, based on clinical signs and symptoms, over the first two weeks postpartum. 2.2.2. Chart review. We queried the electronic medical record system of the clinic to identify all individuals who had been pregnant between 1/1/2006 and 10/1/2013. These charts were individually reviewed to obtain patient age, race, ethnicity, outcome of pregnancy (elective abortion, miscarriage, perinatal demise, or live birth), delivery or termination of pregnancy date, duration of treatment episode, discharge date, and reason for discharge (transferred to another clinic, moved out of area, incarcerated, relapsed and left treatment, tapered off methadone, cross-tapered to buprenorphine, or lost to follow up with no discharge plan). Treatment discontinuation was defined as stopping treatment with no plan for continued medication assisted treatment; as such, it included the following categories of discharge: incarcerated, relapsed and left treatment, tapered off methadone, or lost to follow up with no discharge plan. The beginning of a treatment episode was defined by the first methadone dosing record prior to (or on, if the patient was already pregnant when she began methadone) the date the pregnancy was first identified. The end of the treatment episode was defined as the last dosing record after which there was a break in dosing of at least 14 days or there were no further dosing records. The study was approved by the University of Cincinnati Institutional Review Board.

2. Materials and methods 2.1. Review of currently published literature We completed a systematic review of the published literature to answer the following questions: 1. For pregnant and or postpartum women with opioid use disorder on MAT with methadone or buprenorphine, what are the discontinuation rates of MAT during pregnancy and in the immediate postpartum period? 2. What interventions improve treatment retention compared to standard care for pregnant and postpartum women on MAT with methadone or buprenorphine? We searched PubMed, MEDLINE, and PsycINFO online databases using the Boolean search query: (“medication assisted treatment” OR methadone OR buprenorphine) AND (pregnan* OR postpartum OR post-partum OR “post partum”) AND (retention OR dropout OR drop-out OR “drop out” OR discontinu* OR attendance OR adherence OR compliance). This query yielded 88 unique citations which were imported into an electronic database (EndNote X5; Thomson Reuters, Philadelphia, PA). Titles and abstracts for each article were screened and articles were excluded if it could be determined from screening that they met one of the following pre-specified, sequential exclusion criteria: 1. Pregnant or postpartum women were not a significant focus of the article (studies for which there was only incidental inclusion of pregnant or postpartum women or studies which focused exclusively on neonatal outcomes were excluded). 2. MAT was not addressed as a treatment modality. 3. Article was a review or guidance article without presentation of new data, or a single case report. 4. Article did not report retention, discontinuation, or drop-out rate(s) of MAT or describe intervention(s) to improve treatment retention in MAT. Forty-one articles could not be definitively excluded based on examination of title and abstract and were retained for full-text review. The reference lists for all articles selected for full-text review were manually searched for relevant citations. These were cross-referenced against our original search results and any additional potentially relevant citations were screened. Seven additional articles were added for full-text review based on manual search. We therefore reviewed a total of 48 full text articles using the same exclusion criteria above. Fifteen articles remained after applying all criteria and were included in this review. 2.2. Retrospective cohort study of methadone treatment discontinuation in pregnant and postpartum women 2.2.1. Clinic description. We conducted a retrospective chart review of a methadone treatment clinic affiliated with the University of Cincinnati. The clinic is not-forprofit but does not accept Medicaid or other public assistance. It has a census of approximately 450 individuals receiving methadone dosing, individual therapy, and group counseling. It has collaborative contacts with the maternal-fetal

2.2.3. Statistics. We calculated demographic summary statistics and unadjusted prenatal and 6-month postpartum methadone discontinuation rates. For individuals who had multiple pregnancies during the trial period, only the first pregnancy for each individual was included for determination of discontinuation rates and other analyses. We conducted a Cox Proportional Hazard linear model for postpartum methadone discontinuation, controlling for maternal age, outcome of pregnancy (live versus non-live births), and number of days in treatment at the clinic prior to delivery. Race was not included as a covariate because the population was almost exclusively Caucasian. Outcome of pregnancy was included as a covariate because it was felt that early miscarriages or elective abortions might not be included in patient records, resulting in a sampling difference between the reported non-live births and live births. Individuals who left treatment prior to delivery were excluded from the postpartum Cox proportional hazard analysis and postpartum individuals who transferred to another clinic, moved from the area, or cross-tapered to buprenorphine were censored. One hundred eighty-seven women were included in the Cox proportional hazards model, of which 14 were censored.

3. Results 3.1. Review of currently published literature Table 1 describes the 15 studies identified through our literature review. These studies included a total of 1125 women. Six studies followed women into the postpartum period (Chappel and Senay, 1973; Crandall et al., 2004; Fischer et al., 1998; Laken and Ager, 1996; Peles and Adelson, 2006; Silverman et al., 2001); none were focused exclusively on postpartum subjects. In addition to methadone or buprenorphine prescription, MAT in all studies included some level of group and individual psychosocial treatment at least weekly. Three studies included women treated with buprenorphine (Fischer et al., 1998; Jones et al., 2005, 2010), comprising less than 10% of the total sample reviewed. All other studies used methadone as the only MAT medication, although one study did include 29 women on a slow release formulation (Fischer et al., 1998). 3.1.1. Discontinuation rates of MAT. Four cohort studies reported discontinuation rates for the total population they examined; among these, DePetrillo and Rice (1995) and McCarthy et al. (2005) reported discontinuation rates prior to delivery of 0 and 4% respectively. Two cohort studies included pregnancy and the postpartum period: Peles and Adelson (2006) reported a 1-year discontinuation rate of 22%, and Chappel and Senay (1973) reported a 2-year

Table 1 Studies reporting treatment retention results for MAT in pregnant women. Location

Mean age

Racial composition

Mean EGA (wks) at study entrya

MAT medication and dosage information

Discontinuation rates and other treatment attendance results

n = 133

30.0

16.1

Methadone, mean dose at delivery = 81.3 mg

Overall: 23% (discontinuation prior to delivery)

n = 175 (methadone = 89, buprenorphine = 86)

28.9

71.4% African American, 26.3% Caucasian, 2.3% Biracial 83% Caucasian, 14% African American, 3% Other

18.7

Methadone (51%), mean dose at delivery = 82.9 mg Buprenorphine (49%), mean dose at delivery = 17.2 mg

Overall: 25% (discontinuation prior to delivery) Methadone group: 18% (n = 16) Buprenorphine group: 33% (n = 28)

n = 30 (methadone = 15, buprenorphine = 15)

30

75% African American, 20% Caucasian, 5% Other

23.2

Methadone (50%), mean dose at delivery = 79.1 mg Buprenorphine (50%), mean dose at delivery = 18.7 mg

Overall: 33% (discontinuation prior to delivery) Methadone group: 27% (n = 4) Buprenorphine group 40% (n = 6)

Jones et al. (2001)

n = 85 (intervention = 47; control = 38)

28

76% African American

23.4

Methadone, mean dose = 42 mg

Silverman et al. (2001)

n = 40 (intervention = 20, control = 20)

31.8

83% African American, 17% Caucasian

NR

Methadone, mean dose = 55.5 mg

Svikis et al. (1997) b

n = 66 randomized among 4 treatment groups c

28.3

80.3% African American

22.5

Methadone, dosing NR

Overall: 6% (discontinuation within 14 days) Intervention group: 6.4% (n = 3) Control group: 5.3% (n = 2) Among individuals who did not drop out, the intervention group attended a mean of 12.1 days versus the control group which attended a mean of 10.6 days (p < 0.05) Overall: 53% (discontinuation within 6 months) Intervention group: 45% Control group: 60% Mean treatment duration was 18.6 wks for intervention group and 15.1 wks for control group (p = 0.17) MAT participants: 13.6% (discontinuation within 30 days)

Randomized controlled trials Tuten et al. (2012) Johns Hopkins Center for Addiction and Pregnancy, Baltimore, MD Jones et al. (2010)

Jones et al. (2005)

4 US cities, 2 US rural sites, and Vienna, Austria

Johns Hopkins Center for Addiction and Pregnancy, Baltimore, MD

Cohort studies Peles and Adelson (2006)

Tel Aviv, Israel

n = 45 pregnant women (out of total n = 470 for entire cohort)

31.5

78.3% Israeli, 21.7% Immigrant

NR

Methadone, mean dose at end of study period = 141.1

McCarthy et al. (2005)

Sacramento, CA

n = 94

32

NR

Methadone, mean dose at delivery = 101 mg

Laken et al. (1997) b

Eleonore Hutzel Recovery Program, Detroit, MI

n = 40

29.7

64% Caucasian, 25% Hispanic, 6% African American, 4% Asian, 1% Other 88% African American

26.2

Methadone, dosing NR

n = 55

29.6

26.1

Methadone, dosing NR

Laken and Ager (1996) b

88% African American

C. Wilder et al. / Drug and Alcohol Dependence 149 (2015) 225–231

Sample size

Pregnant women: 22.2% (discontinuation within 1 year; this was not significantly different from the dropout rate of non-pregnant women or of men) Overall 4% (discontinuation prior to delivery) 2% had unavailable outcome information 24.4% attended 4–7 treatment visits; 23.2% attended 8–14, 25.6% attended 15–26, and 26.8% attended 27–96. 44.0% of participants attended no treatment visits; 18.8% attended 1–5 treatment visits; 17.8% attended 6–12 visits, and 20.4% attended 13–62 visits 227

228

Table 1 (Continued) Sample size

Mean age

Racial composition

Mean EGA (wks) at study entrya

MAT medication and dosage information

Discontinuation rates and other treatment attendance results

DePetrillo and Rice (1995)

Location not identified

n = 45

29.3

10.6

Methadone, mean dose at delivery = 52 mg

Overall 0% (discontinuation prior to delivery)

Chappel and Senay (1973)

Special Treatment Unit, Illinois Drug Abuse Program, Chicago, IL

n = 11

NR

78% Caucasian, 22% Latin or African American NR

NR

Methadone, dosing NR

Overall: 63.6% (discontinuation within 2 years)

Hennepin Faculty Associates Addiction Medicine Program, Minneapolis, MN

n = 102 (pregnant cases = 51, non-pregnant controls = 51)

29.9

51% Caucasian, 45% African American, 4% Other

NR

Methadone, dosing NR

Pregnant women: 25.5% (discontinuation within 9 months) Average length of participation was 7.7 months out of a maximum of 9 months which was not significantly different from the control group of non-pregnant women

Johns Hopkins Center for Addiction and Pregnancy, Baltimore, MD

n = 106

30.6

78% African American, 22% Caucasian

14.7

Methadone, mean dose at treatment day 30 = 64 mg

University of Vienna Drug Addiction Outpatient Clinic, Vienna, Austria

n = 98

NR

NR

20.1

Methadone (52%), mean dose at delivery = 45 mg Slow release methadone (35%), mean dose at delivery = 259.4 mg Buprenorphine (12%), mean dose at delivery = 6.6 mg

Average number of days that counseling sessions were attended was 57 for individuals with co-occurring anxiety disorder versus 45 for individuals with either a co-occurring mood disorder or no co-occurring disorder, out of a maximum of 84 days (p < 0.01) Overall: 0% (discontinuation prior to termination of pregnancy or delivery)

Case control studies Crandall et al. (2004)

Observational studies Fitzsimons et al. (2007)

Fischer et al. (1998)

Abbreviations – EGA: estimated gestational age; Wks: weeks; NR: not reported. a Because many studies specified a particular EGA as an entry requirement, the average EGA at study entry may not be representative of the general treatment population. b Only some of the participants in these studies were in MAT (46% of participants in Svikis et al. (1997); 24% in Laken and Ager (1996); and 23.8% in Laken et al. (1997)). Results reported are only for individuals in MAT; in all these studies, methadone was the only MAT medication. c Individuals were randomized to 1 of 4 incentive groups independent of their MAT status.

C. Wilder et al. / Drug and Alcohol Dependence 149 (2015) 225–231

Location

C. Wilder et al. / Drug and Alcohol Dependence 149 (2015) 225–231

3.1.2. Interventions to improve MAT retention. There have been two randomized trials studying the effect of voucher-based incentives on treatment attendance for pregnant women in MAT (Jones et al., 2001; Svikis et al., 1997). An additional RCT was designed to evaluate the effect of voucher-based incentives on drug abstinence but included treatment attendance as 1 of 10 primary outcomes (Tuten et al., 2012). No voucher-based incentive treatment had an effect on its primary treatment retention outcome for MAT participants. Jones et al. (2001) found that an escalating voucher schedule for daily treatment attendance increased the mean number of days of treatment attended during the first 14 days of treatment but did not reduce overall drop-out from treatment in the first 14 days. Svikis et al. (1997) found a non-significant trend for longer treatment retention in the 2 highest compensated voucher groups ($5 or $10 per day) versus the two lowest compensated groups (no compensation or $1 per day) but the overall survival curves for treatment retention in MAT patients did not differ by incentive condition. Tuten et al. (2012) found that voucher group status (no voucher, fixed voucher, or escalating voucher) did not predict number of days in treatment, although it is important to note that in this study, vouchers were designed to reinforce negative urine drug tests, not treatment attendance. Two studies of intensified case management as a tool to improve treatment attendance in pregnant substance users included some women in MAT (Laken and Ager, 1996; Laken et al., 1997), although women receiving MAT comprised less than a quarter of total population in these studies. Both studies conducted path analyses on cohorts of women who received a new case management service provided through a federally funded demonstration project. In both studies, the new case management service was found to be one of several independent components that predicted the total number of treatment visits attended during pregnancy. However, the intervention was not randomized, no control group was available, and actual treatment retention, as opposed to number of treatment visits, was not assessed. 3.2. Retrospective cohort study of methadone treatment discontinuation in pregnant and postpartum women 3.2.1. Sample characteristics. Two hundred twenty-nine women reported 251 pregnancies (20 women had 2 pregnancies and 1 woman had 3) during the 94 month study period. The average age

100%

Estimated Probability of Staying in Treatment

discontinuation rate of 64%, although this cohort included only 11 women. Among randomized controlled trials, discontinuation rates prior to delivery ranged from 6 to 33% (Jones et al., 2001, 2005, 2010; Svikis et al., 1997; Tuten et al., 2012). Silverman et al. (2001) reported a 53% discontinuation rate over a 6-month period which spanned pregnancy and the early postpartum period. It is important to note that these discontinuation rates are specific to research studies which limited patients according to pre-specified selection criteria. The discontinuation rate for the total MAT population from which these studies were drawn may or may not be similar. One case control study which included pregnancy and the postpartum period reported a 9-month discontinuation rate of 26% (Crandall et al., 2004); this rate may or may not be reflective of the total MAT population from which cases and controls were drawn. Finally, an observational study reported a 0% discontinuation rate prior to delivery; it was unclear what proportion of the total population this represented (Fischer et al., 1998). A second observational study did not report discontinuation rates, but did note that pregnant women with co-occurring anxiety disorders attended an average of 57 out of 84 days of counseling sessions compared to an average of 45 out of 84 days for those with either a co-occurring mood disorder or no co-occurring disorder (Fitzsimons et al., 2007).

229

+ Initial n = 187 + 80% + 60%

+ + +

40%

+ = censored (n = 14)

++ 20%

+

0% 0

12

24

36

48

60

72

Months Postpartum Fig. 1. Kaplan–Meier estimates for remaining in methadone treatment after pregnancy.

at delivery or discharge (if prior to delivery) was 28.3 years (SD 4.5). Two hundred twenty-four women (97.8%) were Caucasian, 3 (1.3%) were biracial (Caucasian and African American), and 2 (0.9%) were African American. Six women (2.6%) reported Hispanic ethnicity. Of the 251 reported pregnancies, there were 178 live births, 7 elective abortions, 19 miscarriages, and 3 perinatal deaths reported. The outcomes of 44 pregnancies were unknown because the women left treatment prior to delivery. The number of days of methadone treatment prior to delivery ranged from 0 to 1303, with a median of 145 days and a mean of 233.9 days (SD 265.7). 3.2.2. Prenatal treatment discontinuation. Forty-two patients (18.3%) left the clinic prior to delivery. Of these, 14 transferred to another MAT program, 3 moved out of state, 6 were incarcerated, and 19 were lost to follow-up with no discharge plan. Based on our definition of treatment discontinuation, the prenatal methadone discontinuation rate of this population was 11.0%. 3.2.3. Postpartum treatment discontinuation. The unadjusted methadone discontinuation rate at 6 months postpartum was 61.9%. Fig. 1 shows the Kaplan–Meier estimates for the probability of remaining in methadone treatment following delivery or termination of pregnancy. The estimated probability of methadone discontinuation at or before 6 months postpartum, controlling for age, pregnancy outcome, and duration of treatment prior to delivery, was 56.0%. In the Cox Proportional Hazard linear model, there was no significant relationship between age (HR 0.99, 95% CI (0.95, 1.03)) or pregnancy outcome (HR 1.36, 95% CI (0.86, 2.17)) and postpartum treatment duration. There was a significant inverse relationship between the number of days of clinic treatment prior to delivery and the risk for postpartum discontinuation of methadone (HR 0.98, 95% CI (0.96, 0.99)); that is, the longer the duration of clinic treatment prior to delivery, the lower the likelihood of postpartum methadone discontinuation at any given time. The per-day effect was minimal but cumulative, so, for example, the hazard ratio for receipt of 3 months of pre-delivery methadone treatment was 0.93 (95% CI (0.90, 0.96)), 6 months of pre-delivery methadone treatment was 0.87 (95% CI (0.83, 0.90)), and 12 months of pre-delivery methadone treatment was 0.75 (95% CI = (0.74, 0.76)). 4. Discussion Our review of the published literature suggests that little is definitely known about treatment discontinuation for MAT during pregnancy and the postpartum period, either with respect to general discontinuation rates or interventions to improve treatment retention. Generalizations are not possible based on the limited data available. Although the reviewed studies included a

230

C. Wilder et al. / Drug and Alcohol Dependence 149 (2015) 225–231

total of over 1000 participants, the majority these were involved in clinical trials, where discontinuation was from the study, not from general treatment. Only 4 studies, including a total of just 195 women, included a cohort of treatment recipients potentially generalizable to routine clinical care (Chappel and Senay, 1973; DePetrillo and Rice, 1995; McCarthy et al., 2005; Peles and Adelson, 2006). Among these, discontinuation rates were evaluated at different time points (delivery, 1 year from enrollment, and 2 years from enrollment), making comparison difficult. Our retrospective cohort of 251 pregnancies occurring in routine clinical care more than doubles the number of MAT pregnancies assessed in the published literature. Like most published studies, our cohort included only women on methadone, facilitating comparison of our findings with previous results. We found a prenatal treatment discontinuation rate of 11%, somewhat higher than the 0–4% reported in two previous cohort studies (DePetrillo and Rice, 1995; McCarthy et al., 2005). Perhaps surprisingly, although the Treatment Episode Data Set (TEDS) collected yearly by the Substance Abuse and Mental Health Services Administration (SAMHSA) includes information about average length of stay for MAT facilities (Substance Abuse and Mental Health Services Administration, 2014a), it does not provide specific information about pregnant patients. Its counterpart, the SAMHSA National Survey of Substance Abuse Treatment Services (N-SSATS) database, reports on the number of MAT facilities providing treatment for pregnant patients (Substance Abuse and Mental Health Services Administration, 2014b) but does not provide information on treatment retention. Information on MAT retention for pregnant women at either a national or facility level is not widely available, suggesting the need for studies focused particularly on this topic. Published information on postpartum treatment discontinuation rates for MAT is even scarcer. Of the studies which included both the pregnant and postpartum periods, discontinuation rates varied from 22% to 64%. In our cohort, the probability of discontinuing treatment within 6 months postpartum was 56%, which is within the range of discontinuation rates previously reported in the literature. In our patient sample, the early postpartum period was a time of substantially higher risk for drop-out than the prenatal period. The causes for this are likely multifactorial; in particular, there are probably societal, clinic-specific, and individual reasons driving women to leave treatment after delivery. During the prenatal period, there is considerable medical and societal focus on facilitating MAT for the benefit of the mother and fetus. After delivery however, assistance available to pregnant women frequently ends within a few months; Medicaid for a postpartum woman, for example, may be available for as little as 2 months after the birth of her child. MAT clinics often have clinic-specific incentives to encourage pregnant women to stay in treatment, which may also end after delivery. For instance, in the study clinic, pregnant women qualify for discounted treatment fees but postpartum women do not. Because this clinic does not accept Medicaid, the increase in fees is probably responsible for at least some of the rapid drop-out seen in this cohort. Approximately 35% of methadone clinics nationwide do not accept Medicaid (Substance Abuse and Mental Health Services Administration, 2014b); our results are likely more representative of these clinics. Although there is no published information on discontinuation rates specifically in clinics that accept Medicaid or other public funding, their postpartum dropout rates may be lower than seen in this study. Finally, it is important to remember that individual stressors increase with the birth of a child. Coping with the needs of an infant, often with limited family support, as well as managing increased financial and time constraints are possible reasons that new mothers may leave treatment. Developing interventions to improve treatment retention, particularly during the postpartum period, is critical for improving

treatment retention rates. Unfortunately, our review of the literature also suggests that few studies have tested such interventions. Moreover, voucher-based incentive treatment, the only intervention which has been tested specifically for pregnant women in MAT, was not effective in reducing treatment discontinuation in the three studies which examined this outcome. Despite these discouraging preliminary results, complete dismissal of voucher-based incentives is premature for several reasons. Problems with the design of at least one of the published interventions limit the conclusions that can be drawn (Tuten et al., 2012); no studies have been conducted on voucher-based interventions in conjunction with buprenorphine; and no studies have focused exclusively on the postpartum period. There is a substantial opportunity to improve treatment retention in this patient population. Our results tentatively suggest that longer duration of prenatal methadone treatment may protect against early drop-out in the postpartum period. More rigorous testing of this hypothesis is warranted: interventions to encourage this group of high-risk women to enter methadone treatment as soon as possible in early pregnancy may be one way to improve postpartum methadone retention. Given how little is known about the factors that determine patient drop-out, multiple innovative strategies for treatment retention need to developed, encouraged, and evaluated. There are clear limitations to our current study. Regarding our literature review, it is possible that our search terms were not inclusive enough to identify all articles covering MAT discontinuation in pregnancy. Although we included non-English articles in our review, the databases we searched were primarily English language based, so additional articles not published in English language journals may have been missed. Nonetheless, this review, while not exhaustive, was reasonably comprehensive and reveals a major gap in the research literature worthy of further study. Regarding our cohort study, our findings are strictly descriptive and therefore causal inferences cannot be drawn. We believe a larger population based cohort study is indicated to more accurately determine MAT discontinuation rates postpartum. Rates may vary substantially by geographical region, treatment type (buprenorphine versus methadone), and insurance status. Comparison of fee-for-service MAT clinics with those that accept Medicaid or receive government funding is of particular interest, since the causes for discontinuation may be very different in the two types of clinic. Encouraging all MAT clinics to monitor treatment retention for their postpartum patients is also important. Qualitative or questionnaire based studies to better understand why women leave MAT while pregnant and in the ensuing postpartum period would provide a theoretical basis for developing interventions to improve retention. Opioid use, and the concurrent use of MAT, continues to increase. It behooves us to increase focus on the critical population of pregnant and postpartum women, in which treatment discontinuation affects not only the outcome of the patient, but of her child as well. Our literature review reveals limited information about treatment discontinuation or interventions to improve treatment retention in MAT for pregnant and postpartum women. Results from our cohort study suggest that drop-out rates after delivery are substantially higher than during pregnancy and may be associated with duration of pre-delivery MAT. Current treatment, innovative interventions, and research studies thus far have tended to focus on pregnancy, often ending simultaneous to, or shortly after, delivery. Additional attention to the postpartum period might improve treatment retention and treatment outcomes for new mothers, potentially benefiting families long-term.

Role of funding source Nothing declared.

C. Wilder et al. / Drug and Alcohol Dependence 149 (2015) 225–231

Contributors Christine Wilder, Daniel Lewis, and Theresa Winhusen have reviewed and approved the manuscript and contributed substantially to its content. Conflict of interest No conflict declared. References American Congress of Obstetricians and Gynecologists, 2012. ACOG Committee Opinion No. 524: opioid abuse, dependence, and addiction in pregnancy. Obstet Gynecol 119, 1070–1076, http://dx.doi.org/10.1097/AOG.0b013e318256496e. Bao, Y., Liu, Z., Epstein, D., Du, C., Shi, J., Lu, L., 2009. A meta-analysis of retention in methadone maintenance by dose and dosing strategy. Am. J. Drug Alcohol Abuse 35, 28–33, http://dx.doi.org/10.1080/00952990802342899. Brady, T., Salvucci, S., Sverdlov, L., Male, A., Kyeyune, H., Sikali, E., DeSale, S., Yu, P., 2005. Methadone dosage and retention: an examination of the 60 mg/day threshold. J. Addict. Dis. 24, 23–47. Burns, L., Mattick, R., Lim, K., Wallace, C., 2007. Methadone in pregnancy: treatment retention and neonatal outcomes. Addiction 102, 264–270. Chappel, J., Senay, E., 1973. Special treatment problems in a drug abuse program. Proc. Natl. Conf. Methadone Treat. 2, 1098–1104. Cox, J., Allard, R., Maurais, E., Haley, N., Small, C., 2013. Predictors of methadone program non-retention for opioid analgesic dependent patients. J. Subst. Abuse Treat. 44, 52–60, http://dx.doi.org/10.1016/j.jsat.2012.03.002. Crandall, C., Crosby, R., Carlson, G., 2004. Does pregnancy affect outcome of methadone maintenance treatment? J. Subst. Abuse Treat. 26, 295–303. Dakof, G., Quille, T., Tejeda, M., Alberga, L., Bandstra, E., Szapocznik, J., 2003. Enrolling and retaining mothers of substance-exposed infants in drug abuse treatment. J. Consult. Clin. Psychol. 71, 764–772. Dansky, B., Byrne, C., Brady, K., 1999. Intimate violence and post-traumatic stress disorder among individuals with cocaine dependence. Am. J. Drug Alcohol Abuse 25, 257–268. Deck, D., Carlson, M., 2005. Retention in publicly funded methadone maintenance treatment in two Western States. J. Behav. Health Serv. Res. 32, 43–60. DePetrillo, P., Rice, J., 1995. Methadone dosing and pregnancy: impact on program compliance. Int. J. Addict. 30, 207–217. Faggiano, F., Vigna-Taglianti, F., Versino, E., Lemma, P., 2003. Methadone maintenance at different dosages for opioid dependence. Cochrane Database Syst. Rev., ACD002208, http://dx.doi.org/10.001002/14651858.CD14002208. Fajemirokun-Odudeyi, O., Sinha, C., Tutty, S., Pairaudeau, P., Armstrong, D., Phillips, T., Lindow, S.W., 2006. Pregnancy outcome in women who use opiates. Eur. J. Obstet. Gynecol. Reprod. Biol. 126, 170–175, http://dx.doi.org/10.1016/ j.ejogrb.2005.08.010. Fischer, G., Eder, H., Jagsch, R., Lennkh, C., Habeler, A., Aschauer, H., et al., 1998. Maintenance therapy with synthetic opioids within a multidisciplinary program – a stabilizing necessity for pregnant opioid dependent women. Arch Womens Ment Health. 1, 109–116, http://dx.doi.org/10.1007/s007370050014. Fitzsimons, H., Tuten, M., Vaidya, V., Jones, H., 2007. Mood disorders affect drug treatment success of drug-dependent pregnant women. J. Subst. Abuse Treat. 32, 19–25. Greenfield, L., Brady, J., Besteman, K., De Smet, A., 1996. Patient retention in mobile and fixed-site methadone maintenance treatment. Drug Alcohol Depend. 42, 125–131. Haller, D., Knisely, J., Elswick, R., Dawson, K., Schnoll, S.H., 1997. Perinatal substance abusers: factors influencing retention. J. Subst. Abuse Treat. 14, 513–519. Holbrook, A., Kaltenbach, K., 2012. Co-occurring psychiatric symptoms in opioid-dependent women: the prevalence of antenatal and postnatal depression. Am. J. Drug Alcohol Abuse 38, 575–579, http://dx.doi.org/10.3109/ 00952990.2012.696168. Hulse, G., Milne, E., English, D., Holman, C., 1998. Assessing the relationship between maternal opiate use and neonatal mortality. Addiction 93, 1033–1042. Jones, H., Haug, N., Silverman, K., Stitzer, M., Svikis, D.S., 2001. The effectiveness of incentives in enhancing treatment attendance and drug abstinence in methadone-maintained pregnant women. Drug Alcohol Depend. 61, 297–306, http://dx.doi.org/10.1016/S0376-8716(00)00152-6. Jones, H., Heil, S., Baewert, A., Arria, A., Kaltenbach, K., Martin, P., et al., 2012. Buprenorphine treatment of opioid-dependent pregnant women: a comprehensive review. Addiction 107 (Suppl. 1), 5–27, http://dx.doi.org/10.1111/ j.1360-0443.2012.04035.x. Jones, H., Johnson, R., Jasinski, D., O’Grady, K., Chisholm, C., Choo, R., Crocetti, M., Dudas, R., Harrow, C., Huestis, M.A., Jansson, L.M., Lantz, M., Lester, B.M., Milio, L., 2005. Buprenorphine versus methadone in the treatment of pregnant

231

opioid-dependent patients: effects on the neonatal abstinence syndrome. Drug Alcohol Depend. 79, 1–10, http://dx.doi.org/10.1016/j.drugalcdep.2004.11.013. Jones, H., Kaltenbach, K., Heil, S., Stine, S., Coyle, M., Arria, A., O’Grady, K.E., Selby, P., Martin, P.R., Fischer, G., 2010. Neonatal abstinence syndrome after methadone or buprenorphine exposure. N. Engl. J. Med. 363, 23202331, http://dx.doi.org/10.1056/NEJMoa1005359. Jones, H., Svikis, D., Tran, G., 2002. Patient compliance and maternal/infant outcomes in pregnant drug-using women. Subst. Use Misuse 37, 1411–1422. Kandall, S., Doberczak, T., Jantunen, M., Stein, J., 1999. The methadone-maintained pregnancy. Clin. Perinatol. 26, 173–183. Kelly, S., O’Grady, K., Mitchell, S., Brown, B., Schwartz, R., 2011. Predictors of methadone treatment retention from a multi-site study: a survival analysis. Drug Alcohol Depend. 117, 170–175, http://dx.doi.org/10.1016/ j.drugalcdep.2011.01.008. Kissin, W., Svikis, D., Morgan, G., Haug, N., 2001. Characterizing pregnant drugdependent women in treatment and their children. J. Subst. Abuse Treat. 21, 27–34. Laken, M., Ager, J., 1996. Effects of case management on retention in prenatal substance abuse treatment. Am. J. Drug Alcohol Abuse 22, 439–448. Laken, M.P., McComish, J.F., Ager, J., 1997. Predictors of prenatal substance use and birth weight during outpatient treatment. J. Subst. Abuse Treat. 14, 359–366, http://dx.doi.org/10.1016/s07405472(96)00006-2. Ludlow, J., Evans, S., Hulse, G., 2004. Obstetric and perinatal outcomes in pregnancies associated with illicit substance abuse. Aust. N. Z. J. Obstet. Gynaecol. 44, 302–306, http://dx.doi.org/10.1111/j.1479-828X.2004.00221.x. MacGowan, R., Swanson, N., Brackbill, R., Rugg, D., Barker, T., Molde, S., 1996. Retention in methadone maintenance treatment programs, Connecticut and Massachusetts, 1990–1993. J. Psychoactive Drugs 28, 259–265. Magura, S., Nwakeze, P., Demsky, S., 1998. Pre-and in-treatment predictors of retention in methadone treatment using survival analysis. Addiction 93, 51–60. Mattick, R., Breen, C., Kimber, J., Davoli, M., 2009. Methadone maintenance therapy versus no opioid replacement therapy for opioid dependence. Cochrane Database Syst. Rev., http://dx.doi.org/10.1002/14651858.CD002209.pub2. McCarthy, J., Leamon, M., Parr, M., Anania, B., 2005. High-dose methadone maintenance in pregnancy: maternal and neonatal outcomes. Am. J. Obstet. Gynecol. 193, 606–610, http://dx.doi.org/10.1016/j.ajog.2005.03.072. McGuire, F., 1979. Arrest record and retention in methadone maintenance treatment. Int. J. Addict. 14, 1009–1013. Minozzi, S., Amato, L., Bellisario, C., Ferri, M., Davoli, M., 2013. Maintenance agonist treatments for opiate-dependent pregnant women. Cochrane Database Syst. Rev. 12, CD006318, http://dx.doi.org/10.1002/14651858.CD006318.pub3. Newman, R., Tytun, A., Bashkow, S., 1976. Retention of patients in the New York City methadone maintenance treatment program. Int. J. Addict. 11, 905–931. Peles, E., Adelson, M., 2006. Gender differences and pregnant women in a methadone maintenance treatment (MMT) clinic. J. Addict. Dis. 25, 39–45. Peles, E., Schreiber, S., Bloch, M., Dollberg, S., Adelson, M., 2012. Duration of methadone maintenance treatment during pregnancy and pregnancy outcome parameters in women with opiate addiction. J. Addict. Med. 6, 18–23, http://dx.doi.org/10.1097/ADM.0b013e318229bb25. Silverman, K., Svikis, D., Robles, E., Stitzer, M., Bigelow, G., 2001. A reinforcementbased therapeutic workplace for the treatment of drug abuse: six-month abstinence outcomes. Exp. Clin. Psychopharmacol. 9, 14–23. Substance Abuse and Mental Health Services Administration, 2003. Results from the 2002 National Survey on Drug Use and Health: National Findings. NHSDA Series H-22, HHS Publication No. (SMA) 03-3836. Substance Abuse and Mental Health Services Administration, Rockville, MD. Substance Abuse and Mental Health Services Administration, 2013. Results from the 2012 National Survey on Drug Use and Health: National Findings. NHSDA Series H-46, HHS Publication No. (SMA) 13-4795. Substance Abuse and Mental Health Services Administration, Rockville, MD. Substance Abuse and Mental Health Services Administration, 2014a. Treatment Episode Data Set (TEDS): 2011. Discharges from Substance Abuse Treatment Services. BHSIS Series S-70, HHS Publication No. (SMA) 14-4846. Substance Abuse and Mental Health Services Administration, Rockville, MD. Substance Abuse and Mental Health Services Administration, 2014b. 2011 Opioid Treatment Program Survey: Data on Substance Abuse Treatment Facilities with OTPs. BHSIS Series S-65, HHS Publication No. (SMA) 14-4807. Substance Abuse and Mental Health Services Administration, Rockville, MD. Svikis, D., Lee, J., Haug, N., Stitzer, M., 1997. Attendance incentives for outpatient treatment: effects in methadone-and nonmethadone-maintained pregnant drug dependent women. Drug Alcohol Depend. 48, 33–41. Tuten, M., Svikis, D., Keyser-Marcus, L., O’Grady, K., Jones, H., 2012. Lessons learned from a randomized trial of fixed and escalating contingency management schedules in opioid-dependent pregnant women. Am. J. Drug Alcohol Abuse 38, 286–292, http://dx.doi.org/10.3109/00952990.2011.643977. Welle-Strand, G.K., Skurtveit, S., Jones, H.E., Waal, H., Bakstad, B., Bjarko, L., Ravndal, E., 2013. Neonatal outcomes following in utero exposure to methadone or buprenorphine: a new National Cohort Study of opioid-agonist treatment of pregnant women in Norway from 1996 to 2009. Drug Alcohol Depend. 127, 200–206.