Cannabis Legalization and Cannabis-Involved Pregnancy Hospitalizations in Colorado

Abstract

The primary objective of this study was to evaluate the association between presence of recreational cannabis dispensaries and prevalence of cannabis-involved pregnancy hospitalizations in Colorado. This was a retrospective cohort study of pregnancy-related hospitalizations co-coded with cannabis diagnosis codes in the Colorado Hospital Association from January 1, 2011, through December 31, 2018 (recreational cannabis began January 1, 2014). Our primary outcome was cannabis-involved pregnancy hospitalizations per 10k live births per county. The primary exposure measure was county variation in the number of recreational dispensaries. We controlled for counties’ baseline exposure to medical cannabis dispensaries and used Poisson regression to evaluate the association between exposure to recreational cannabis and hospitalizations. During the study period, cannabis-involved pregnancy hospitalizations increased from 429 to 1,210. Mean hospitalizations per county (1.7 to 4.7) and per 10k live births (13.2 to 55.7) increased. Overall, increasing recreational dispensaries were associated with increases in hospitalizations (1.02, CI: 1.00,1.04). When comparing counties with different densities of baseline medical cannabis market, low and high exposure counties had fewer hospitalizations than those counties with no exposure (low: IRR 0.97, CI: 0.96–0.99; high: 0.98, CI: 0.96–0.99). In Colorado, there was more than a two-fold increase in cannabis-involved pregnancy hospitalizations between 2011 and 2018. Counties with no baseline exposure to medical cannabis had a greater increase than other counties, suggesting the recreational market may influence cannabis use among pregnant individuals.

INTRODUCTION:

Cannabis is one of the most used recreational drugs reported during pregnancy around the world.[1–4] Canada has legalized cannabis as a country, while over half of US states have allowed medical cannabis and almost 20% have allowed legalized recreational cannabis. In the US, approximately 1 in 20 pregnant individuals report cannabis use at some time during pregnancy,[5] Reported use has increased among this population in the US in the past decade.[6] Surveillance data from California, Colorado and Washington, US states that have legalized both medical and recreational cannabis, demonstrated an increase in cannabis use or detection in pregnant individuals associated with cannabis legalization.[7–12] Another concerning trend is not only recreational cannabis use during pregnancy, but also self-treatment of pregnancy-related side effects, including nausea and vomiting.[13, 14] The lack of regulation of cannabis industry marketing claims may be contributing to increases in cannabis use to self-treat pregnancy-related conditions. For example, one study found that 69% of marijuana dispensaries in Colorado recommended cannabis use for first-trimester pregnancy related nausea and vomiting.[15]

Research has found that prenatal cannabis use may adversely affect fetal growth and neurodevelopment, be associated with future learning and behavioral problems in children, and increase risk of neonatal morbidity.[16–30] Several US organizations, including the American College of Obstetricians and Gynecologists and the American Academy of Pediatrics, recommend abstinence from cannabis use during pregnancy.[29, 31] Despite these recommendations, there are concerns that the trend of cannabis use during pregnancy will continue to increase as more states and countries legalize cannabis and allow for commercial sales through dispensaries.[32]

When cannabis was legalized in Colorado, local jurisdictions were granted the authority to allow either medical or recreational cannabis dispensaries or both. This resulted in county-level variation in cannabis availability due to both markets, which provides an opportunity to examine if differences in number and type of dispensaries in a local market are associated with pregnancy inpatient hospitalizations with cannabis-use diagnostic codes. The objective of this study was to examine the trend over time in cannabis-involved pregnancy hospitalizations, and to evaluate the association between density of recreational cannabis dispensaries at the county-level and cannabis-involved hospitalizations Colorado. We hypothesized that counties with existing large medical dispensary markets would already be exposed to more general cannabis use in the local population, and thus show less of an effect on our main outcome than in counties without any baseline medical dispensary exposure.

METHODS

This was a retrospective observational study using two Colorado statewide administrative data sources for the period January 1, 2011, through December 31, 2018. The Colorado Department of Revenue (DOR) oversees the state’s cannabis industry. Colorado began accepting medical cannabis dispensary applications in 2010. Amendment 64, which allowed sale of recreational cannabis passed on November 6, 2012. The sale of recreational cannabis began in January 2014 in existing medical cannabis dispensaries, and in independent recreational cannabis dispensaries in September 2014. Our primary outcome was cannabis-involved pregnancy hospitalizations, including labor and delivery, reported to the Colorado Hospital Association (CHA) per county per 10k live births. ICD-9 and ICD-10 diagnosis codes (ICD9: V22*-V28*, V91* 630–679; ICD10: Z33-Z37*, P961*, P962*, O*) were used to identify pregnancy-related hospitalizations and cannabis use or dependence claims (ICD9: 304.3, 305.2; ICD10: F12*) claims from the CHA database. Claims were aggregated to the county level per quarter by patient ZIP code. We used the number of live births per county per year as the population adjustment for each county, which we obtained from the Colorado Department of Public Health and Environment.[33]

We used two sources of cannabis exposure. The main exposure was the number of recreational dispensaries per county per quarter. We also needed to control for county exposure to medical dispensaries since these were legalized before recreational dispensaries. The DOR provided dispensary location (name, address) and date of licensure for all licensed dispensaries (medical or recreational). We categorized counties in terms of the relative size of the medical market prior to legalization (no baseline medical exposure = 0 medical dispensaries; low baseline = 1–9 medical dispensaries; high baseline = 10+ medical dispensaries). We then interacted this variable with the number of recreational dispensaries.

We used descriptive statistics to examine trends of our outcome during the study period. Models, which are estimated in STATA 16.1 using the xtpoisson command, also include controls for the point at which diagnosis coding switched to ICD version 10, county-level unemployment, and the total number of hospitalizations, as well as quarter, year and county fixed effects. The unit of analysis is the county-quarter. An interrupted time series analysis was performed using Stata’s itsa command running a linear model using the data aggregated across all counties. The RAND Internal Review Board exempted this research from human subjects’ protection.

RESULTS

From January 1, 2011, through December 31, 2018, there were a total of 6,229 pregnancy-related hospitalizations co-coded with cannabis reported to CHA. The most common age group distribution in our cohort was individuals aged 19 to 25 years (49.0%,), followed by individuals 26 years and older (44.0%), then 18 years and younger (7.1%). The most common insurance payer was government (84.7%), followed by commercial (12.3%), and other types of insurance (0.9%). Statewide, the number of cannabis-involved pregnancy hospitalizations increased from 429 in 2011 to 1,210 in 2018 (Table 1). This increase was most notable after 2012 and 2014 when accounting for average visits per county and per 10k live births (Figure 1).

Table 1:

Annual Cannabis Dispensary Descriptive Statistics and Cannabis-Involved Pregnancy Hospitalizations in Colorado, 2011 through 2018.

	Full Period 2011–2018 (n=2,048)	2011 (n=256)	2012 (n=256)	2013 (n=256)	2014 (n=256)	2015 (n=256)	2016 (n=256)	2017 (n=256)	2018 (n=256)
Med Dispensaries per County, Mean (Min, Max)	8.0 (0.0, 221.0)	7.8 (0.0, 209.0)	7.9 (0.0, 209.0)	8.1 (0.0, 221.0)	7.9 (0.0, 208.0)	8.2 (0.0, 212.0)	8.3 (0.0, 215.0)	8.1 (0.0, 211.0)	7.8 (0.0, 202.0)
Med Dispensaries per County per 10k Residents, Mean (Min, Max)	1.0 (0.0, 10.8)	1.0 (0.0, 9.2)	1.1 (0.0, 9.2)	1.1 (0.0, 10.8)	1.1 (0.0, 7.0)	1.0 (0.0, 7.6)	0.9 (0.0, 8.1)	0.9 (0.0, 5.3)	0.8 (0.0, 5.2)
Rec Dispensaries per County, Mean (Min, Max)	4.1 (0.0, 186.0)	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)	3.6 (0.0, 131.0)	6.0 (0.0, 159.0)	7.0 (0.0, 176.0)	7.8 (0.0, 185.0)	8.5 (0.0, 186.0)
Rec Dispensaries per County per 10k Residents, Mean (Min, Max)	1.2 (0.0, 39.4)	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)	0.0 (0.0, 0.0)	0.8 (0.0, 8.8)	1.9 (0.0, 29.1)	2.3 (0.0, 29.0)	2.4 (0.0, 28.1)	2.6 (0.0, 39.4)
Total (Med+Rec) per County, Mean (Min, Max)	12.1 (0.0, 394.0)	7.8 (0.0, 209.0)	7.9 (0.0, 209.0)	8.1 (0.0, 221.0)	11.5 (0.0, 337.0)	14.1 (0.0, 371.0)	15.3 (0.0, 389.0)	15.9 (0.0, 394.0)	16.3 (0.0, 386.0)
Cannabis-involved pregnancy hospitalizations per County, Mean (Min, Max)	3.0 (0.0, 48.0)	1.7 (0.0, 24.0)	1.5 (0.0, 23.0)	2.0 (0.0, 32.0)	2.9 (0.0, 44.0)	3.6 (0.0, 42.0)	3.8 (0.0, 38.0)	4.2 (0.0, 48.0)	4.7 (0.0, 48.0)
Cannabis-involved pregnancy hospitalizations per 10k live births per County, Mean (Min, Max)	35.1 (0.0, 689.7)	13.2 (0.0, 384.6)	13.4 (0.0, 400.0)	22.1 (0.0, 555.6)	36.7 (0.0, 625.0)	45.2 (0.0, 666.7)	45.5 (0.0, 487.8)	49.3 (0.0, 689.7)	55.7 (0.0, 666.7)
Cannabis-involved pregnancy hospitalizations, Statewide, N	6229	429	396	519	735	909	960	1071	1210

Min=minimum, Max=maximum. Med = medical; Rec = Recreational. Colorado has 64 counties, and the unit of analysis is the county-quarter. The means reflect the average across the 64 counties and four quarters in the calendar year (264 observations per year).

Figure 1. Cannabis-Involved Pregnancy Hospitalizations Reported to the Colorado Hospital Association, 2011–2018.

Note: The vertical line at quarter 1 of 2014 marks the start of recreational cannabis sales in Colorado. We classified counties according to their baseline level of medical dispensary exposure (no medical dispensaries; 1–9; and 10+) in 2012. The right panel is adjusted for the number of live births per county per year.

The average number of medical dispensaries remained constant throughout the study period (Table 1), while the number of recreational dispensaries grew quickly after 2014, accounting for most of the overall growth of the cannabis dispensary market. A rise in cannabis-involved pregnancy hospitalizations also increased after recreational dispensary sales began, more than doubling in mean claims per county (1.7 to 4.7) between 2011 to 2018 and quadrupling in mean claims (13.2 to 55.7) per 10,000 live births.

Our regression results in Table 2 show that the number of recreational dispensaries was positively associated with the outcome, although small (IRR 1.02, 95% CI: 1.00, 1.04). When comparing counties with different densities of baseline medical cannabis market, counties with low and high levels of baseline exposure had fewer pregnancy claims co-coded with cannabis than those counties with no baseline exposure (low: IRR 0.97, CI: 0.96–0.99; high: 0.98, CI: 0.96–0.99). This was consistent with our hypothesis that counties with no exposure at baseline would see more effects from recreational marijuana.

Table 2:

Impact of Cannabis Dispensaries on Cannabis-Involved Pregnancy Hospitalizations per 10K Live Births, 2011–2018.

Impact of Recreational Dispensaries on Cannabis-Involved Pregnancy Hospitalizations	IRR [95% CI]
Impact of recreational dispensary count	1.02* [1.00,1.04]
Impact of recreational dispensary count on counties with low number of baseline medical dispensaries	0.97** [0.96,0.99]
Impact of recreational dispensary count for counties with high number of baseline medical dispensaries	0.98* [0.96,0.99]
Time Variables
Seasonal Quarter (Quarter 1, reference)
Quarter 2	1.00 [0.95,1.06]
Quarter 3	1.10** [1.03,1.18]
Quarter 4	1.13** [1.04,1.23]
Year (2011, reference)
2012	1.00 [0.88,1.13]
2013	1.50** [1.10,2.04]
2014	3 29*** [1.92,5.63]
2015	5 12*** [2.74,9.56]
2016	5.98***
	[2.95,12.10]
2017	7 78*** [3.53,17.15]
2018	8 44*** [3.90,18.27]
Post ICD10 switch indicator (=1 in Quarter 3 2015)	1.03 [0.87,1.22]
County unemployment rate	1.19* [1.04,1.37]
Total hospital admissions	1.00+ [1.00,1.00]
N	1888
Combined IRR for the Impact of Recreational Dispensaries on Counties with Different Baseline Medical Cannabis Dispensary Densities.
No medical dispensary exposure	1.02 (p=0.042)
Low medical dispensary exposure (1–10)	0.994 (p=0.11)
High medical dispensary exposure (>10)	0.995 (p=0.000)

Notes: IRR = incidence rate ratio; CI = confidence interval. We classified counties according to their baseline level of medical dispensary exposure (no medical dispensaries; 1–9; and 10+) in 2012.

^***Indicates significance at the 0.01% level

^**significance at the 1% level

^*significance at the 5% level, and

⁺indicates significance at 10% level. All models estimated using xtpoisson in STATA 16.1 with robust standard errors, using the number of live births per county per year as the exposure. The 1,888 observations reflects the number of county-quarters included in the analysis. Colorado has 64 counties, but five counties (148 county-quarter observations) drop out of the analysis because there are no pregnancy-related claims co-coded with cannabis.

DISCUSSION

Pregnancy-related complications or diagnoses are not approved health conditions for medical cannabis in Colorado, so the prevalence of the medical cannabis market should not have affected the prevalence of use in this population. However, legalizing recreational cannabis allowed any adult, including pregnant individuals, to purchase cannabis without oversight. Thus, we would expect the legalization of recreational cannabis to be associated with an increase in use of cannabis by pregnant people at the population level, which we do in fact observe in statewide cannabis-involved pregnancy hospitalizations. We also observed counties without medical cannabis dispensaries had greater increases in hospitalizations post recreational legalization. This potentially indicates there may have been an impact of the new recreational market on pregnant individuals’ use decisions.

Other states, such as Washington and California, have also observed an increase in self-reported cannabis use and/or positive toxicology screens during the prenatal period after cannabis legalization.[7, 8, 10, 11, 13] According to the National Survey on Drug Use and Health (NSDUH), past month cannabis use increased from 3.4% to 7.0% among pregnant individuals between 2002 to 2017, a period in which 8 states legalized recreational cannabis. [6] Frequency of use among pregnant women also rose. This increase in cannabis use among pregnant individuals was most evident among those in their first trimester, increasing from 5.7% to 12.1%. Using the 2016 Pregnancy Risk Assessment Monitoring system (PRAMS), Skelton and colleagues found that people delivering a live-born infant in three states that had legalized recreational cannabis were significantly more likely to use cannabis during the preconception, prenatal and postpartum period compared to those in three states without legalized recreational cannabis.[34] In the 4.2% of individuals who self-reported cannabis use during pregnancy in the 2017 PRAMS, the most common reasons for cannabis use were to self-treat and relieve stress or anxiety (81.5%), nausea or vomiting (77.8%), and pain (55.1%).[14]

Cannabis use during pregnancy has been associated with several at-risk socioeconomic factors, which may contribute to worse outcomes for children: younger age, lower income, unemployment, lack of education, and being a victim of abuse.[29, 35, 36] Individuals who use cannabis during pregnancy have higher rates of alcohol, tobacco, and other substance use.[29, 35, 36] Children who have caregivers who have substance use are at higher risk for child maltreatment, child welfare evaluations and out-of-home placements for their children.[37–44]

In addition to socioeconomic risks, there are growing concerns for the biological effects on the fetus from prenatal cannabis use, including preterm delivery, decreased birthweight and other growth parameters.[16–22, 24–29, 45, 46] Additional research has observed an association between prenatal cannabis exposure and adverse childhood development including decreased capacity for verbal reasoning, memory, language skills, and an increase in impulsivity and hyperactivity. However, some of these published studies have limitations in their methodological approach, it is difficult to account for all confounders, and limited information exists on long-term physiologic and neurodevelopmental consequences.[30]

With increasing legalization in the US and elsewhere, additional surveillance and research is needed to determine the short- and long-term effects of modern high potency cannabis products on pregnancy, perinatal, and subsequent childhood outcomes. The commercialized cannabis industry has quickly grown, providing a variety of concentrated products for public purchase.[47, 48] Concentrated products include tetrahydrocannabinol-infused food products (edibles) and concentrates (dabs, budders, waxes) for vaping.[49] Although smoking remains the most common route of use in pregnancy, other modes of use are recently being observed. Reproductive-age females from 8 states participating in PRAMS reported alternative modes of cannabis use including eating (12.1%), vaporizing (7.1%), dabbing (4.5%) and drinking (0.5%).[10] Most studies evaluating the health effects of cannabis use, including in the pregnancy population, do not consider nor study the impact of use of these highly concentrated products.

As a result of these concerns for both pregnant individuals and their children, identification of use during pregnancy can provide an opportunity for targeted interventions. However, many pregnant individuals are reluctant to admit use and seek assistance, even when residing in a state that has legalized cannabis, due to stigma and fear of persecution. Discrepancies in federal and state laws can still lead to mandatory reporting and involvement of Child Protective Services in many states from healthcare providers since cannabis remains a federal Schedule I drug.[32] Going forward, considerations of new policies and screening approaches are needed to determine how best to accurately recognize and screen for cannabis use during pregnancy, and provide effective interventions that benefit the maternal-infant dyad.

STUDY STRENGTHS AND LIMITATIONS:

We used Colorado as our study setting, a state in the US that has legalized medical and recreational cannabis. Our results may not apply to other states and countries as timelines of implementation of sales and regulations of the cannabis industry may vary. Our study was unique in using local cannabis dispensaries as a marker of cannabis availability and access for consumers, including the pregnant population. Using the number of local dispensaries allowed us to evaluate the impact of greater access and availability of cannabis on the study population. CHA data represents most acute care hospitals in the state (90 of the 108 hospitals), though there is some variation in reporting over time.[50] Unlike other data sources, patient claims in CHA are a representation of the entire state and includes uninsured residents. However, there are some limitations to our analysis. First, we identified inpatient healthcare encounters using both pregnancy and cannabis ICD diagnosis codes. Documentation of a cannabis-involved diagnosis could represent a variety of reasons and severity of cannabis use in relation to their healthcare encounter. The cannabis use may have been significant enough to be directly related to the encounter, such as intoxication or cannabis use disorder. Additional reasons for the encounter could include history of use, or a positive toxicology screen. We were unable to do a detailed medical record abstraction to determine how cannabis specifically related to the inpatient admission, or during what trimester of their pregnancy the hospital visit occurred. Improvements in accuracy of coding and standardization in testing and reporting of prenatal care patients could enhance surveillance methodology. Because of non-standardization of cannabis screening within healthcare facilities, our results are likely an underrepresentation of the true number of cannabis-related pregnancy inpatient encounters. In Colorado, cannabis use in pregnancy remains reportable to social services as it is a federal schedule I drug despite recreational legalization at the state level. Previous data demonstrated an initial increase in reporting which subsequently decreased presumably due to recognition of ongoing reporting requirements for practitioners.[11, 32]

CONCLUSION:

In Colorado, there was more than a two-fold increase in cannabis-involved pregnancy hospitalizations between 2011 and 2018. This increase was highest after sale of recreational cannabis began in 2014. We observed that counties with no baseline exposure to medical cannabis had a greater increase than other counties, suggesting the presence of the recreational market may influence cannabis use decisions among pregnant individuals. There are concerns of the impact of cannabis use on the maternal-infant dyad, and specifically on the neurodevelopment of the child. With increasing legalization and decriminalization in the US and other countries around the world, public health and state officials may need to develop tighter regulations and public education to limit use during pregnancy and be able to more effectively and safely provide resources to support cannabis cessation in this population.

HIGHLIGHTS:

• There was more than a two-fold increase in cannabis-involved pregnancy hospitalizations

• There was a four-fold increase in the rate of pregnancy hospitalizations per 10K live births.

• This increase was highest after sale of recreational cannabis began in 2014.

• The expansion of recreational markets was associated with an increase in cannabis-involved pregnancy hospitalizations.

• Counties with no medical markets prior to recreational legalization experienced greater rises than other counties.