A nation-wide Swedish study of opioid analgesic prescribing patterns during pregnancy and associated pre-existing mental health conditions

Abstract

Introduction:

Research has consistently shown individuals with mental health conditions are more likely to be prescribed opioid analgesic medications and to engage in heavier utilization. However, it is unclear whether these findings apply to pregnant women.

Study design:

We explored opioid analgesic prescription in 689,400 pregnancies occurring in Sweden between 2007 and 2013. We investigated prescription patterns across time and type of source clinic for any opioid analgesic and for strong and weak opioid analgesics. We further evaluated the extent to which receipt of opioid analgesic medications was associated with previous mental health diagnoses and prescriptions of other psychoactive medications.

Results:

The prevalence of pregnant women who filled prescriptions for opioid analgesics (4.5%) was relatively stable across the assessed years. However, among pregnant women who filled opioid analgesic prescriptions, there was a large increase in strong opioid analgesic prescriptions–from 6.1% in 2007 to 17.1% in 2013. The main source of opioid analgesic prescriptions were primary care and obstetrics and gynecology clinics–38.7% of all filled prescriptions originated from primary care providers and 25.3% from obstetrics and gynecology practitioners. Compared to pregnant women who did not fill any opioid analgesic prescriptions, those who did were more likely to have a wide range of pre-existing mental health diagnoses (e.g., anxiety disorder odds ratio[OR]=3.13, 95% confidence interval[CI]:2.98,3.29) and to utilize a wide range of other psychoactive medications (e.g., benzodiazepines OR=4.26, 95% CI:4.10,4.43). Similarly, those who received strong opioids were more likely to have a wide range of mental health diagnoses and be prescribed a wide range of psychoactive medications compared to those who received weak opioids.

Conclusions:

These results highlight the need for physicians treating pregnant women and women of childbearing age for painful conditions to obtain detailed histories of mental health problems, screen for symptoms of mental health problems, and facilitate integrated care and evidence-based mental health interventions if needed.

Introduction

Research has suggested that women are prescribed opioids more than men[1]. Women may be in greater need of pain management due to common painful conditions that predominantly affect women, such as fibromyalgia[2], endometriosis[3] and uterine fibroid tumors[4]. Therefore, there have been major calls for studies of prescription opioid use in women, and these calls have emphasized the need for research among women of reproductive age and pregnant women[5]. Understanding prescription opioid use during pregnancy is particularly important because maternal health and behavior during pregnancy could affect the course of pregnancy and, thereby, the health of both the mother and the developing offspring.

Research assessing opioid analgesic use during pregnancy is sparse and existing findings are mixed. The estimated prevalence of opioid analgesic use during pregnancy varies widely across samples and years[e.g., 6, 7]. Although studies of general patient populations have suggested that individuals prescribed opioids may be particularly in need of mental health services by consistently showing individuals with pre-existing mental health conditions and psychoactive medication use are more likely to be prescribed opioids and engage in heavier utilization (e.g., longer duration of use and higher doses)[e.g., 8, 9–11], research specifically evaluating these associations among pregnant women is limited. Some studies have shown that pregnant women with mental health conditions are more likely to be prescribed opioids[12, 13]; however, previous studies used relatively small samples and did not compared the prevalence of mental health conditions among pregnant women with lighter versus heavier opioid utilization. Moreover, research is especially lacking on co-occurring use of other psychoactive medications, which may be prescribed to treat comorbid mental health conditions.

The present study aimed to address gaps in our understanding of opioid analgesic prescribing patterns during pregnancy and mental health conditions among pregnant women prescribed an opioid analgesic, particularly pregnant women specifically prescribed strong opioid analgesics. First, we examined opioid analgesic prescribing patterns across (a)time and (b)types of prescribing clinic. Second, we evaluated mental health conditions among pregnant women prescribed any opioid analgesic, as well as pregnant women specifically prescribed strong opioid analgesics. In addition to studying associations with pre-existing mental health diagnoses, we also assessed associations with filled prescriptions of other psychoactive medications during pregnancy.

Methods

A STROBE checklist (Table S1) documents that the necessary items were included in this report of a cohort study.

This population-based, observational study used data recorded in national Swedish registers and included all singleton pregnancies without missing data that resulted in live births between January, 1^st 2007 and December, 31^st 2013. We linked information from seven Swedish registers (Table S2)[14–20].

Measures

Demographic characteristic

Assessed demographic characteristics included age at childbearing, highest level of education in year of delivery, family income in year of delivery (average of maternal and paternal income), neighborhood deprivation score in year of delivery, country of origin, and county of residence in year of delivery. Following previous research[21], neighborhood deprivation scores incorporated annual proportions of welfare recipients, unemployed individuals, immigrants, divorced individuals, and individuals with low educational attainment and measures of residential mobility, crime rates, and neighborhood disposable income.

Opioid analgesic prescriptions

We identified prescriptions of opioid analgesics filled during the period between conception and delivery. Consistent with prior research[22], we first estimated last menstrual period by subtracting gestational age (predominantly based on ultrasound measurements in the 18^th-20th week of pregnancy[23]) from delivery date and then estimated conception date by adding 14 days to last menstrual period.

We defined opioid analgesic prescription as medications with anatomical therapeutic class codes beginning with N02A, as well as other paracetamol/codeine combinations and buprenorphine and methadone prescribed as analgesics (Table S3 shows specific medications). We distinguished buprenorphine and methadone for pain treatment versus opioid use disorder treatment with criteria used in previous publications[24]. We also identified prescriptions of strong opioid analgesics (i.e., all opioid analgesics except codeine, dextropropoxyphene, and tramadol) and type of clinic from which the prescriptions originated.

Mental health conditions and co-occurring psychoactive medication

We identified several mental health conditions diagnosed before conception according to the International Classification of Disease 10^th edition criteria codes (Table S4) and prescriptions of several other classes of psychoactive medications filled between conception and delivery (Table S5). Prior research has validated mental health diagnoses in the Swedish register data[25, 26].We additionally categorized women as having no, one, or more than one diagnosed condition(s) and categorized women as having none, one class, or more than one class of co-occurring medication.

Statistical analyses

First, to characterize the sample, we evaluated demographic characteristics of the analytic sample. Second, to evaluate opioid analgesic prescription patterns, we assessed yearly prevalence of pregnancies with filled opioid analgesic prescriptions and documented types of clinics the prescriptions originated from. Third, to evaluate mental health conditions among pregnant women prescribed any opioid analgesic, we assessed the prevalence of mental health diagnoses and utilization of other psychoactive mediations according to opioid analgesic prescription status. We used odds ratios (ORs) to quantify the associations between these characteristics and (a)receipt of any opioid analgesic in the entire analytic sample and (b)receipt of strong opioids among all opioid analgesic recipients. We obtained ORs and their 95% confidence intervals (CIs) from logistic regression models that included indicators for year to account for potential time trends.

Sensitivity analysis

First, given that we only had diagnoses made by specialists before pregnancy to assess pre-existing mental health conditions, and, therefore, may not have captured less severe conditions, we evaluated associations with prescriptions of psychoactive medications filled before pregnancy (i.e., in the 360 days before pregnancy) in order to capture women who may have symptoms without receiving diagnoses from specialists. Second, to further specify the type of opioid analgesic, we evaluated associations between mental health diagnoses and psychoactive mediations during pregnancy with strong, long-lasting opioids compared to weak long-lasting opioids (i.e., tramadol) and immediate-release opioids. To classify opioids as immediate release versus long lasting, we used information on product names and route of administration (i.e., medications administered via injections or rectally were considered immediate release).

Ethics approval

This study used data from national Swedish registers and was approved by the institutional review board at Indiana University and the regional ethical review board in Stockholm, Sweden. By Swedish law, informed consent was unnecessary.

Results

Sample

To create the analytic sample, we started with all 765,318 births in the Medical Birth Register between January, 1^st 2007 and December, 31^st 2013 and sequentially excluded births with invalid maternal identifiers (117), stillbirths (2,829), multiple births (21,285), and births missing data on gestational age (200), offspring sex (1), or demographic characteristics (51,486). Table 1 presents demographics of the resulting sample of 689,400 pregnancies, which represents 90.1% of the study base.

Table 1.

Demographics of the study sample

	n (%)
Maternal age at childbearing
<20 years	5900 (1.4)
20 to 29 years	265932 (76.4)
30 to 39 years	383290 (11.4)
40 to 45 years	33400 (10.5)
>45 years	878 (0.3)
Maternal highest level of education by 2013
< 9 years	21158 (5.9)
9 years	55690 (20.2)
1 to 3 years upper secondary	260432 (81.2)
any post-secondary or postgraduate	352120 (92.6)
Family income at year of delivery
1st quintile	49417 (13.9)
2nd quintile	86732 (26.3)
3rd quintile	204675 (60.6)
4th quintile	237218 (68.2)
5th quintile	111358 (31.0)
Neighborhood deprivation score at year of delivery
1st quintile	103101 (28.9)
2nd quintile	120951 (34.5)
3rd quintile	125162 (36)
4th quintile	147179 (42.9)
5th quintile	193007 (57.7)
Maternal Swedish country of origin	538867 (158.2)
Geographic area
Malmo	93366 (24.2)
Stockholm	173244 (53.8)
Goteborg	116228 (33.1)
Other counties	306563 (88.9)

Prescription patterns

Across the entire cohort of 689,400 pregnancies, 30,427 (4.4%) had opioid analgesic prescriptions, and among these, 3,202 (10.5%) were prescribed strong opioids. The prevalence of pregnancies with opioid analgesic prescriptions steadily rose from 3.9% in 2007 to 4.9% in 2010, then dropped to approximately 4.4% in 2011 and remained relatively stable for the following two years (Figure S1). The drop in 2011 corresponds to the removal of the opioid dextropropoxyphene from the market in Sweden. In contrast, prescription of strong opioid analgesics consistently increased from 6.1% of all opioid prescriptions in 2007 to 17.1% in 2013 (Figure 1). This translates into a 0.2% to 0.8% increase among all pregnant women. The most common sources of the prescriptions were primary care (38.7%) and obstetrics and gynecology clinics (25.3), and this pattern was largely constant across the assessed years (Table S6).

Figure 1.

Changes in the prevalence of strong opioid analgesic prescription

Of note, the prevalence of opioid analgesic prescriptions was not higher among pregnancies excluded due to missing demographic data (3.9%), indicating that excluding these pregnancies did not result in a sample that was biased towards including more severe cases.

Mental health conditions associated with receipt of any opioid analgesic prescription

Compared to all other pregnant women, women who filled a prescription for any opioid analgesic during pregnancy were more likely to have each of the assessed pre-existing mental health diagnoses (Table S7). The most common co-occurring mental health diagnosis was opioid use disorder (OR=4.90, 95% CI:4.03,5.96) followed by anxiety disorders (OR=3.13, 95% CI:2.98,3.29). Pregnant women with opioid analgesic prescriptions were also more likely to have multiple mental health diagnoses (OR=3.5, 95% CI:2.16,2.35). Additionally, receipt of any opioid analgesic prescription in pregnancy was associated with higher odds of receiving other psychoactive medications during pregnancy (Table S7), particularly tricyclic antidepressants (OR=9.78, 95% CI:8.42,11.36), and being prescribed multiple classes of psychoactive mediations during pregnancy (OR=5.71, 95% CI:5.42,6.02).

Mental health conditions associated with receipt of strong opioid analgesic prescription

Women who received strong opioid analgesics during pregnancy were no different from women who received weak opioid analgesic prescriptions with respect to pre-existing diagnoses of schizophrenia or bipolar disorder (OR=1.02, 95% CI:0.71,1.45), but they were more likely to have diagnoses of all other assessed mental health conditions (Table S7). Similar to any opioid prescriptions, among strong opioid analgesic recipients, the most common diagnosed co-morbid condition was opioid use disorder (OR=3.11, 95% CI:2.09,4.65) followed by anxiety disorders (OR=1.49, 95% CI:1.30,1.70). Those who received strong opioids were also more likely to have multiple mental health diagnoses compared to those who received weak opioids (OR=1.56, 95% CI:1.34,1.81).

Recipients of strong opioid analgesics were not more likely to be prescribed non-benzodiazepine anxiolytics (OR=1.08, 95% CI:0.77,1.51) or non-tricyclic antidepressants (OR=1.01, 95% CI:0.88,1.16) during pregnancy compared to women receiving weak opioid analgesics. In contrast, women who received strong opioid analgesics had eleven times the odds of being prescribed an opioid use disorder medication during pregnancy as compared to those who received weak opioid analgesics (OR=11.16, 95% CI:6.37,19.58). We observed moderate differences in prescribing rates for the other assessed psychoactive medications. Women with strong opioid analgesic prescriptions were also more likely to be prescribed psychoactive medications from multiple other classes during pregnancy compared to women with weak opioids prescriptions (OR=1.34, 95% CI:1.16,1.54).

Sensitivity analyses

Consistent with the findings for associations with pre-existing mental health diagnoses, pregnant women with opioid analgesic prescriptions were more likely to have filled prescriptions for psychoactive medication before pregnancy as compared to pregnant women without opioid prescriptions; and, pregnant women with strong opioid analgesic prescriptions were more likely to have filled prescriptions for psychoactive medication before pregnancy as compared to pregnant women with weak opioid analgesic prescriptions (Table S8).

The pattern of results with strong, long-lasting opioid analgesics versus other opioids was also similar to the pattern of results with strong versus weak opioid analgesics, though associations were often larger in analyses with strong, long-lasting medications (Table S9).

Discussion

In a sample of 689,400 women giving birth in Sweden between 2007 and 2013, 4.5% filled prescriptions for opioid analgesics during the pregnancy. Among those with filled opioid analgesic prescriptions, strong opioids became more common over the assessed years. Our findings expand upon previous research that has documented that individuals in the general population with mental health conditions are more likely to be prescribed opioids and engage in heavier use[e.g., 8, 9, 10] by showing that these results also apply to pregnant women. Specifically, we found that pregnant women prescribed opioid analgesics were more likely to have pre-existing mental health conditions and be prescribed other psychoactive medications compared to all other pregnant women. We also found pre-existing mental health conditions and prescriptions for other psychoactive medications were more common among pregnant women receiving strong compared to weak opioid analgesics.

Our results have important clinical implications. Physicians treating pregnant women should be aware that utilization of strong opioid analgesics in pregnancy appear to have increased, and pregnant women prescribed opioid analgesics, particularly those with strong opioid prescriptions, are likely to have co-occurring mental health conditions and be prescribed other psychoactive medications. These results highlight the increasing need for physicians treating pregnant women with painful conditions to obtain a detailed mental health history, including a history of diagnoses of mental health conditions and use of psychoactive medications. Given that mental health conditions often are undiagnosed, physicians who do not typically conduct mental health screenings should screen pregnant women seeking treatment for painful conditions for symptoms of mental health problems. Arguably, physicians should obtain a mental health history and screen all women of childbearing age, not just pregnant women, with painful conditions given that many pregnancies are unplanned. Recent advances in computer adaptive testing and remote assessments may help expedite the screening of mental health problems, including mental health problems among perinatal women[27, 28]. If physicians detect conditions that appear to be unmanaged, physicians should either directly address the mental health problems with evidence-based interventions (e.g., brief evidence-based psychological interventions)[29] or make referrals to physicians that offer evidence-based treatment of mental health problems. Physicians should also communicate any additional information they gathered from mental health histories and screenings to other providers in an effort to coordinate care (e.g., management of multiple medications during pregnancy). It may be particularly important for physicians in primary care and obstetrics and gynecology to be trained in screening and brief interventions given that our data showed that opioid analgesic prescriptions tended to originate from primary care and obstetrics and gynecology clinics.

More studies are needed to further inform clinical practice by addressing unanswered research questions and limitations of our study. First, based on the research to date it is unclear what factors could explain why mental health conditions and opioid analgesic prescriptions co-occur. There are multiple potential explanations including (a)genetic, neural, and psychological mechanisms, such as sensitivity to reward and stress, causing both indications for analgesia and mental health conditions[30–32], (b)mental health condition making individuals more vulnerable to pain and, therefore, in need of more pain management treatment, (c)patients with mental health conditions seeking out opioid analgesic prescriptions not only to manage pain but also to manage psychological conditions, and (d)physicians diagnosing patients with mental health conditions, particularly opioid use disorder, based on a patient history of chronic opioid prescriptions. Second, it is unclear what factors are responsible for the increase in utilization of strong opioid analgesics among pregnant women over time. For example, it would be helpful to know if the increase in utilization of strong opioid analgesics corresponds to an increase in prevalence of conditions causing severe pain. Third, it would be helpful to gain knowledge on other important predictors of opioid analgesic prescriptions during pregnancy, such as physical health conditions. Fourth, studies are needed to evaluate associations between opioid analgesic prescriptions during pregnancy and continuous measures of mental health in order to capture a range of symptom severity. While we had mental health diagnoses made by specialists, we did not have mental health diagnoses made by general practitioners or continuous measures of mental health symptoms. Therefore, we may have only captured severe mental health conditions and missed mild to moderate conditions. However, we observed a similar pattern of results when examining associations with mental health diagnoses made by specialists as when examining associations with prescriptions of psychoactive medications filled before pregnancy, which is a different measure of the same general construct and arguably captures less severe mental health conditions. Moreover, research conducted on the general population found the same pattern of results when using Swedish data including only specialist diagnoses[9] as when using US commercial health insurance data that captured diagnosed conditions with a wider range of symptom severity[10]. Nonetheless, more detailed screenings of mental health problems in future studies could facilitate an understanding of milder and more common mental health issues among pregnant women who use opioid analgesic medications. Fifth, similar analyses to ours conducted in other samples would help evaluate whether our findings generalize to other populations.

Conclusion

Our study showed that in Sweden the utilization of strong opioid analgesics among pregnant women increased over time, most opioid analgesic prescriptions made during pregnancy originated from primary care and obstetrics and gynecology clinics, and pregnant women prescribed opioid analgesics, particulary strong opioids, were at increased risk for having mental health conditions. These results highlight the need for physicians treating pregnant women and women of childbearing age for painful conditions to obtain detailed histories of mental health problems and screen for symptoms of mental health problems. If physicians detect unmanaged conditions, they should facilitate integrated care and evidence-based mental health interventions.