Maternal Substance Abuse and the Later Risk of Fractures in Offspring: L’abus maternel de substances et le risque ultérieur de fractures chez les enfants

Abstract

Objective:

To assess the association of maternal illicit drug abuse before or during pregnancy with future fractures in offspring.

Methods:

We performed a longitudinal cohort study of 792,022 infants born in hospitals of Quebec, Canada, between 2006 and 2016, with 5,457,634 person-years of follow-up. The main exposure was maternal substance abuse before or during pregnancy, including cocaine, opioid, cannabis, and other illicit drugs. The main outcome measure was hospitalization for traumatic fracture in offspring up to 12 years of age. We used adjusted Cox regression models to compute hazard ratios (HR) and 95% confidence intervals (CI) for the association of maternal drug abuse with the subsequent risk of fracture in children.

Results:

The incidence of child fractures was higher for maternal illicit drug abuse than no drug abuse (21.2 vs. 15.4 per 10,000 person-years). Maternal drug abuse before or during pregnancy was associated with 2.35 times the risk of assault-related fractures (95% CI, 1.29 to 4.27) and 2.21 times the risk of transport accident-related fractures (95% CI, 1.34 to 3.66), compared with no drug abuse. Associations were strongest before 6 months of age for assault-related fractures (HR = 2.14; 95% CI, 0.97 to 4.72) and after 6 years for transport-related fractures (HR = 2.86; 95% CI, 1.35 to 6.05). Compared with no drug abuse, associations with assault and transport-related fractures were elevated for all drugs including cocaine, opioids, and cannabis.

Conclusions:

Maternal illicit drug abuse is associated with future child fractures due to assault and transport accidents.

Introduction

Substance use is prevalent in women of reproductive age,^1–3 but implications for future fractures in their children receive limited attention. Some data indicate that parental substance use is associated with an increased risk of pediatric injuries,⁴ but the association with fractures has not been studied. Fractures are among the most common preventable injuries,^5,6 costing Canadians nearly $27 billion annually.⁷ Fractures occur in 27% to 64% of children and are a major cause of morbidity as they account for nearly 60% of hospitalized traumatic injuries before 18 years of age.^5,6 Moreover, prevention is challenging due to the paucity of data on the determinants of pediatric fractures, including maternal drug abuse. Maternal illicit drug abuse may lead to suboptimal parenting or child maltreatment, factors that can affect the risk of fracture.^4,8

The relationship between maternal drug abuse and risk of childhood fractures is not known although there is hypothesis generating research for alcohol.^9,10 Previous studies suggest that maternal alcohol use is associated with 3 times the risk of fracture in children under 18 years⁹ and 2 times the odds of long-bone fractures before 5 years of age.¹⁰ No study has investigated the relationship with maternal illicit drug abuse. Many women initiate illicit drug use early in adolescence and continue to use drugs during pregnancy.^2,11,12 More than a third of high school girls in the United States report using marijuana and 1% to 7% trying drugs such as cocaine and heroin at least once.¹¹ Early onset substance use is a risk factor for continued use.¹³ As substance use disorders are highly undertreated relative to other psychiatric conditions,¹⁴ prenatal substance use could be a marker for continued use in women of childbearing age and potentially a marker of risk to their offspring.

In the United States, the prevalence of prenatal cannabis use doubled between 2002 and 2014,² and rates of opioid use increased 4-fold between 1999 and 2014.¹² Determining whether illicit drug use early in the lives of women is associated with the future risk of fractures in offspring may provide an upstream window of opportunity for fracture prevention. We assessed the association between maternal substance abuse before or during pregnancy with the future risk of traumatic fracture in offspring.

Methods

Study Design and Participants

We conducted a population-based longitudinal cohort study of 792,022 neonates born between 2006 and 2016 in all hospitals of Quebec, Canada. We followed infants over time from birth until the end of the study on March 31, 2018, to identify traumatic fractures that led to hospitalization during childhood. We obtained de-identified data on each child from the Maintenance and Use of Data for the Study of Hospital Clientele registry.¹⁵ This registry gathers discharge abstracts of hospitalized individuals in Quebec, including the cause of injury and up to 41 diagnostic and 35 procedure codes. Infant hospital charts are linked with the mother. We were able to capture maternal health conditions during pregnancy or that led to hospitalization before pregnancy any time after April 1, 1989.

Infants who died at birth were not eligible for the study nor were infants with missing health insurance numbers who could not be followed over time. We excluded infants with osteogenesis imperfecta, a genetic condition associated with bone fragility and fractures.¹⁶

Measures

Maternal Drug Abuse

The main exposure was maternal drug abuse before or during pregnancy (yes, no), including cocaine, opioid, cannabis, and other illicit drugs (stimulants, hallucinogens, sedatives, hypnotics, and volatile solvents; Supplemental Table S1). In this study, drug abuse encompassed dependence, overdose, and poisoning. Using diagnostic codes from the 9th and 10th revisions of the International Classification of Diseases, we identified women who were hospitalized for problematic drug use any time before pregnancy. We also identified women who used drugs during pregnancy, as documented on their obstetric chart at delivery (Supplemental Table S1). The obstetric chart captures substance use that was self-reported or discovered by the physician through screening.¹⁷ These charts provide information on the entire obstetric history and do not rely on prenatal hospitalizations. We also identified infants with neonatal abstinence syndrome, a complication generally attributed to opioid exposure.¹⁸

Childhood Fractures

The main outcome measure included fractures that were serious enough to require in-patient hospital treatment. We did not analyze minor fractures that were treated in an ambulatory setting. Using diagnostic codes, we identified all infants who sustained a fracture that required hospitalization up to 12 years of age (Supplemental Table S2). We focused on traumatic fractures that occurred any time after the birth hospitalization and did not consider pathological or stress fractures. We classified fractures by cause, including assault, transport accident, fall, mechanical force, and other causes. We further grouped fractures by anatomical location, including head (skull and facial bones); spine, thorax, and pelvis; upper limb (shoulder, humerus, forearm, hand); and lower limb (femur, lower leg, foot). As a measure of severity, we identified children with single versus multiple fractures.

Covariates

We considered several maternal confounders, including age at delivery (<25, 25 to 34, ≥35 years); parity (0, 1, ≥2 previous deliveries); preterm birth (<37- vs. ≥37-week gestation); multiple birth (yes, no); pregnancy complications defined as preeclampsia and preexisting or gestational diabetes (yes, no); mental illness defined as schizophrenia, mood, anxiety, stress, and personality disorders, or suicide attempt (yes, no); and alcohol use disorders (yes, no; Table S1). We accounted for infant sex (male, female) and comorbidity (osteoporosis, other bone density and structure disorders, congenital heart defects, vitamin D, calcium, or phosphorus deficiency, cholestasis, and bronchopulmonary dysplasia; Table S1). We considered socioeconomic deprivation (high, moderate-high, moderate, moderate-low, low, unspecified), place of residence (rural, urban, unspecified), and time period at birth (2006 to 2009, 2010 to 2012, 2013 to 2016) as other potential confounders. Socioeconomic deprivation captured poverty for quintiles of the population based on neighborhood income, education, and employment.¹⁹

Statistical Analyses

We determined the incidence of pediatric fractures per 10,000 person-years. Using the cumulative incidence function which accounts for death as a competing event, we computed the cumulative incidence of fractures after 12 years of follow-up based on type of drug abused (cocaine, opioids, cannabis) and cause of fracture (assault, transport accident, fall). In Cox proportional hazards regression models, we estimated the association of maternal substance abuse with the risk of fracture in children, expressed as hazard ratios (HR) and 95% confidence intervals (CI). Models were adjusted for maternal age, parity, preterm birth, multiple birth, pregnancy complications, mental illness, alcohol use disorders, infant sex, infant comorbidity, socioeconomic deprivation, place of residence, and time period. As the time scale, we used the number of days from birth until the first hospitalization for fracture, death or end of the study on March 31, 2018. We applied the Fine and Gray method to account for death as a competing event.²⁰ Children who were never hospitalized for fractures before the end of the study were censored. We used robust sandwich estimators to account for infants with the same mother.

We assessed the association of maternal substance abuse with the risk of fracture by cause, anatomical site, and severity of the injury. We analyzed the risk of fracture by age (<0.5, 0.5 to 1.4, 1.5 to 2, 3 to 5, ≥6 years) to determine whether the associations varied during childhood and by type of illicit drug abuse. In sensitivity analyses, we assessed the impact of maternal prepregnancy and pregnancy substance abuse separately to determine whether the timing of abuse influenced the future risk of fracture. We also analyzed combined drug and alcohol abuse in relation to fracture risk.

Statistical analyses were carried out in SAS version 9.4 (SAS Institute Inc., Cary, NC). As we analyzed de-identified data, informed patient consent was not required, and we received an ethics waiver from the institutional review board of the University of Montreal Hospital Centre.

Results

In this cohort of 792,022 infants born between 2006 and 2016 with 5,457,634 person-years of follow-up, 12,576 infants (1.6%) had mothers with a drug use disorder before or during pregnancy (Table 1). Among 12,576 exposed infants, 176 sustained at least 1 fracture that required hospitalization before 12 years of age, for an incidence of 21.4 per 10,000 person-years (95% CI, 18.5 to 24.8). In comparison, the incidence of fractures was 15.4 per 10,000 person-years (95% CI, 15.1 to 15.8) for children whose mothers had no illicit drug abuse history.

Table 1.

Incidence of Bone Fracture Requiring Hospitalization according to Maternal and Child Characteristics.

Characteristic	Total No. of Infants	No. of Fractures	Incidence per 10,000 Person-years (95% Confidence Interval)
Substance use disorder
Yes	12,576	176	21.4 (18.5 to 24.8)
No	779,446	8,296	15.4 (15.1 to 15.8)
Age, years
<25	127,183	1,709	19.0 (18.1 to 19.9)
25-34	529,138	5,593	15.3 (14.9 to 15.7)
≥35	135,701	1,170	13.1 (12.4 to 13.9)
Parity at delivery
0	387,541	3,864	14.4 (14.0 to 14.9)
1	277,026	3,060	16.0 (15.5 to 16.6)
≥2	127,455	1,548	17.9 (17.0 to 18.8)
Preterm birth, weeks’ gestation
<37	50,724	625	17.9 (16.5 to 19.4)
≥37	741,298	7,847	15.4 (15.0 to 15.7)
Multiple birth
Yes	12,516	166	20.1 (17.2 to 23.4)
No	779,506	8,306	15.5 (15.1 to 15.8)
Pregnancy complicationsa
Yes	95,813	953	15.4 (14.4 to 16.4)
No	696,209	7,519	15.5 (15.2 to 15.9)
Mental illnessb
Yes	34,889	434	19.4 (17.6 to 21.3)
No	757,133	8,038	15.4 (15.0 to 15.7)
Alcohol use disorder
Yes	4,157	62	22.8 (17.7 to 29.2)
No	787,865	8,410	15.5 (15.2 to 15.8)
Infant sex
Male	406,412	4,802	17.2 (16.7 to 17.6)
Female	385,610	3,670	13.8 (13.4 to 14.3)
Infant comorbidityc
Yes	13,456	225	25.2 (22.1 to 28.7)
No	778,566	8,247	15.4 (15.0 to 15.7)
Socioeconomic deprivation
High	156,580	1,806	16.8 (16.0 to 17.6)
Moderate-high	153,862	1,781	16.8 (16.0 to 17.6)
Moderate	154,747	1,720	16.1 (15.3 to 16.8)
Moderate-low	156,182	1,607	14.8 (14.1 to 15.5)
Low	140,076	1,298	13.2 (12.5 to 14.0)
Place of residence
Rural	145,594	2,078	20.6 (19.7 to 21.5)
Urban	632,495	6,290	14.4 (14.0 to 14.7)
Time period
2006 to 2009	306,310	5,051	16.7 (16.2 to 17.2)
2010 to 2012	244,782	2,405	15.2 (14.6 to 15.8)
2013 to 2016	240,930	1,016	12.0 (11.3 to 12.8)
Total	792,022	8,472	15.5 (15.2 to 15.9)

^a Preeclampsia, preexisting or gestational diabetes.

^b Schizophrenia, mood, anxiety, stress, and personality disorders, or suicide attempt before or during pregnancy.

^c Osteoporosis, other bone density and structure disorders, congenital heart defects, vitamin D, calcium, or phosphorus deficiency, cholestasis, bronchopulmonary dysplasia.

Children whose mothers abused illicit drugs had a higher cumulative incidence of fractures overall (Figure 1). The incidence of fracture hospitalization diverged rapidly in the first months after birth regardless of the type of maternal drug abuse. In children with any maternal drug abuse history, the incidence of assault-related fractures accelerated rapidly in early infancy. However, the incidence of fractures due to transport accidents only began to increase around 4 years of age. The incidence of fall-related fractures did not vary greatly between children with and without a history of maternal drug abuse.

Figure 1.

Maternal drug abuse before or during pregnancy and cumulative incidence of childhood fractures in offspring. Cumulative incidence of childhood fractures by (A) type of drug used and (B) cause of fracture. Solid line = maternal drug abuse, dotted line = no drug abuse.

In adjusted regression models, infants whose mothers had a history of substance abuse had 1.14 times the risk of any childhood fracture (95% CI, 0.97 to 1.35), compared with no maternal drug abuse (Table 2). Maternal drug abuse was associated with 2.35 times the risk of assault-related fractures (95% CI, 1.29 to 4.27) and 2.21 times the risk of fractures due to transport accidents (95% CI, 1.34 to 3.66), but there was no association with other causes. Associations were particularly high for head (HR = 1.55; 95% CI, 1.15 to 2.08); spine, thorax, or pelvis (HR = 2.86; 95% CI, 1.53 to 5.37); shoulder (HR = 3.22; 95% CI, 1.08 to 9.56); and femur fractures (HR = 1.56; 95% CI, 1.05 to 2.31). Maternal drug abuse was also more strongly associated with multiple fractures (HR = 1.85; 95% CI, 1.06 to 3.21) than single fractures (HR = 1.12; 95% CI, 0.94 to 1.33).

Table 2.

Association of Maternal Substance Abuse with Childhood Fractures.

	Maternal Substance Abuse (N = 12,576)		No Substance Abuse (N = 779,446)		Hazard Ratio (95% Confidence Interval)a
Childhood fracture	No. of Factures	Incidence per 10,000 Person-years	No. of Factures	Incidence per 10,000 Person-years	Unadjusted	Adjustedb
Cause of fracture
Any	176	21.4	8,296	15.4	1.40 (1.21 to 1.63)	1.14 (0.97 to 1.35)
Assault	18	2.2	185	0.3	6.06 (3.73 to 9.84)	2.35 (1.29 to 4.27)
Transport accident	24	2.9	549	1.0	2.98 (1.98 to 4.48)	2.21 (1.34 to 3.66)
Fall	98	11.9	6,188	11.5	1.04 (0.86 to 1.27)	0.91 (0.73 to 1.13)
Mechanical force	21	2.5	751	1.4	1.86 (1.21 to 2.87)	1.41 (0.90 to 2.20)
Other	21	2.5	762	1.4	1.79 (1.16 to 2.76)	1.40 (0.87 to 2.24)
Site of fracture
Head	62	7.5	1,905	3.5	2.08 (1.62 to 2.68)	1.55 (1.15 to 2.08)
Skull	42	5.1	1,322	2.4	2.00 (1.47 to 2.72)	1.60 (1.11 to 2.31)
Face	22	2.7	594	1.1	2.46 (1.61 to 3.76)	1.48 (0.91 to 2.42)
Spine, thorax, pelvis	15	1.8	167	0.3	5.78 (3.40 to 9.81)	2.86 (1.53 to 5.37)
Upper limb	81	9.8	4,926	9.1	1.10 (0.89 to 1.37)	0.95 (0.74 to 1.21)
Shoulder	7	0.8	99	0.2	4.54 (2.11 to 9.76)	3.22 (1.08 to 9.56)
Humerus	40	4.8	2,392	4.4	1.11 (0.81 to 1.52)	1.12 (0.79 to 1.60)
Forearm	36	4.4	2,259	4.2	1.08 (0.78 to 1.50)	0.82 (0.57 to 1.16)
Hand	<5	0.5	330	0.6	0.82 (0.31 to 2.18)	0.59 (0.21 to 1.69)
Lower limb	40	4.8	1,523	2.8	1.70 (1.24 to 2.32)	1.31 (0.94 to 1.82)
Femur	29	3.5	960	1.8	1.93 (1.33 to 2.78)	1.56 (1.05 to 2.31)
Lower leg	12	1.5	509	0.9	1.56 (0.88 to 2.76)	1.04 (0.56 to 1.94)
Foot	<5	0.5	96	0.2	2.76 (1.02 to 7.49)	1.78 (0.66 to 4.77)
Severity of fracture
Multiple	18	2.2	377	0.7	3.13 (1.95 to 5.01)	1.85 (1.06 to 3.21)
Single	162	19.7	7,938	14.8	1.35 (1.16 to 1.57)	1.12 (0.94 to 1.33)

^a Hazard ratios above 1 indicate an increased risk of fracture for substance abuse relative to no abuse and are statistically significant when 95% confidence intervals do not cross 1.

^b Adjusted for maternal age, parity, preterm birth, multiple birth, pregnancy complications, mental illness, alcohol use disorders, infant sex, infant comorbidity, socioeconomic deprivation, place of residence, and time period.

In analyses stratified by age, associations appeared to be stronger with fractures early in infancy (Table 3). Compared with no drug abuse history, maternal drug abuse was associated with 1.70 times the risk of any fracture between 0 and 5 months (95% CI, 1.09 to 2.64), an association that appeared to be mainly driven by assault-related fractures (HR = 2.14; 95% CI, 0.97 to 4.72). Statistically significant associations were present for head and upper limb fractures in this age group. Between 6 months and 5 years of age, maternal drug abuse was not associated with childhood fractures, regardless of cause of injury. At 6 years and older, however, the association strengthened, with maternal drug abuse associated with 2.86 times the risk of fractures due transport accidents (95% CI, 1.35 to 6.05).

Table 3.

Association of Maternal Substance Abuse with Childhood Fractures according to Age at Injury.

	Hazard Ratio (95% Confidence Interval)a
Childhood fracture	<6 Months	6 to 17 Months	18 to 35 Months	3 to 5 Years	≥6 Years
Cause of fracture
Any	1.70 (1.09 to 2.64)	1.02 (0.64 to 1.63)	1.32 (0.92 to 1.91)	0.96 (0.70 to 1.31)	1.07 (0.76 to 1.51)
Assault	2.14 (0.97 to 4.72)	2.23 (0.70 to 7.11)	1.20 (0.17 to 8.56)	30.30 (2.98 to 307.62)	—
Transport accident	—	3.07 (0.28 to 33.60)	1.86 (0.44 to 7.94)	1.83 (0.83 to 4.04)	2.86 (1.35 to 6.05)
Fall	1.70 (0.98 to 2.96)	0.66 (0.33 to 1.31)	1.09 (0.64 to 1.84)	0.81 (0.54 to 1.21)	0.80 (0.52 to 1.25)
Mechanical force	1.05 (0.15 to 7.15)	2.08 (0.76 to 5.66)	2.04 (0.90 to 4.61)	0.82 (0.31 to 2.13)	1.48 (0.56 to 3.88)
Other	1.49 (0.41 to 5.42)	1.27 (0.37 to 4.37)	1.41 (0.63 to 3.17)	1.16 (0.38 to 3.59)	1.69 (0.59 to 4.84)
Site of fracture
Head	1.78 (1.02 to 3.11)	0.65 (0.30 to 1.42)	1.90 (1.08 to 3.34)	1.50 (0.72 to 3.11)	2.22 (0.99 to 4.98)
Spine, thorax, pelvis	1.68 (0.52 to 5.48)	1.52 (0.74 to 3.12)	5.55 (0.66 to 46.75)	5.72 (2.74 to 11.91)	2.62 (0.33 to 20.63)
Upper limb	2.51 (1.04 to 6.04)	1.46 (0.53 to 4.05)	1.35 (0.69 to 2.62)	0.66 (0.42 to 1.03)	0.94 (0.63 to 1.41)
Lower limb	1.59 (0.68 to 3.77)	1.20 (0.57 to 2.54)	0.87 (0.44 to 1.73)	2.11 (1.14 to 3.91)	0.84 (0.27 to 2.65)

^a Hazard ratios above 1 indicate an increased risk of fracture for substance abuse relative to no abuse and are statistically significant when 95% confidence intervals do not cross 1. Hazard ratios are adjusted for maternal age, parity, preterm birth, multiple birth, pregnancy complications, mental illness, alcohol use disorders, infant sex, infant comorbidity, socioeconomic deprivation, place of residence, and time period.

All drugs were associated with assault-related fractures, compared with no drug abuse (Table 4). Risks were elevated for cocaine (HR = 4.41; 95% CI, 0.91 to 21.26), opioids (HR = 2.41; 95% CI, 0.55 to 10.63), cannabis (HR = 3.62; 95% CI, 1.52 to 8.66), and other drugs (HR = 2.88; 95% CI, 1.14 to 7.27) although associations were not all statistically significant. Associations with fractures due to transport accidents were also apparent, but associations with other causes of fracture were weaker or absent. When examined by site, opioid abuse was associated with head fractures (HR = 1.95; 95% CI, 1.07 to 3.56), and cocaine abuse was associated with lower limb fractures (HR = 2.92; 95% CI, 1.51 to 5.63).

Table 4.

Association of Specific Types of Drugs with Childhood Fractures.

	Hazard Ratio (95% Confidence Interval)a
Childhood fracture	Cocaine (N = 1,876)	Opioids (N = 1,833)	Cannabis (N = 3,307)	Other Drugsb (N = 3,310)
Cause of fracture
Any	1.29 (0.86 to 1.93)	1.09 (0.73 to 1.63)	1.17 (0.86 to 1.59)	0.96 (0.69 to 1.34)
Assault	4.41 (0.91 to 21.26)	2.41 (0.55 to 10.63)	3.62 (1.52 to 8.66)	2.88 (1.14 to 7.27)
Transport accident	1.68 (0.51 to 5.60)	1.95 (0.60 to 6.35)	2.95 (1.21 to 7.21)	2.11 (0.83 to 5.38)
Fall	1.07 (0.64 to 1.80)	0.81 (0.47 to 1.41)	0.85 (0.56 to 1.29)	0.68 (0.43 to 1.07)
Mechanical force	1.91 (0.73 to 5.01)	1.52 (0.50 to 4.58)	0.86 (0.30 to 2.45)	1.78 (0.78 to 4.03)
Other	1.05 (0.21 to 5.19)	1.51 (0.47 to 4.79)	1.74 (0.80 to 3.78)	0.53 (0.13 to 2.12)
Site of fracture
Head	1.73 (0.87 to 3.46)	1.95 (1.07 to 3.56)	1.53 (0.90 to 2.61)	1.52 (0.88 to 2.61)
Spine, thorax, pelvis	5.12 (0.98 to 26.76)	1.67 (0.22 to 12.54)	2.46 (0.65 to 9.35)	4.86 (1.98 to 11.91)
Upper limb	0.82 (0.43 to 1.58)	0.71 (0.37 to 1.38)	0.80 (0.49 to 1.31)	0.60 (0.35 to 1.03)
Lower limb	2.92 (1.51 to 5.63)	0.75 (0.24 to 2.35)	1.67 (0.94 to 2.97)	1.19 (0.61 to 2.31)
Severity of fracture
Multiple	1.97 (0.47 to 8.23)	—	1.21 (0.33 to 4.38)	2.41 (1.06 to 5.51)
Single	1.24 (0.81 to 1.90)	1.15 (0.77 to 1.73)	1.19 (0.87 to 1.63)	0.90 (0.63 to 1.29)

^a Hazard ratios above 1 indicate an increased risk of fracture for substance abuse relative to no abuse and are statistically significant when 95% confidence intervals do not cross 1. Hazard ratios are adjusted for maternal age, parity, preterm birth, multiple birth, pregnancy complications, mental illness, alcohol use disorders, infant sex, infant comorbidity, socioeconomic deprivation, place of residence, and time period.

^b Stimulants, hallucinogens, sedatives, hypnotics, and volatile solvents.

Sensitivity analyses revealed that drug abuse both during and before pregnancy was strongly associated with assault and transport-related fractures. Prepregnancy drug abuse was associated with 2.08 times the risk of assault-related fractures (95% CI, 1.02 to 4.25) and 2.06 times the risk of transport-related fractures (95% CI, 1.06 to 3.99), compared with no drug abuse history. Pregnancy drug abuse was associated with 2.72 times the risk of assault-related fractures (95% CI, 1.23 to 6.02) and 2.39 times the risk of transport-related fractures (95% CI, 1.25 to 4.54). Comorbid illicit drug and alcohol abuse was associated with a slightly greater risk of assault and transport-related fractures than drug abuse only (Supplemental Table S3).

Discussion

This population-based cohort study of 792,022 children with nearly 5.5 million person-years of follow-up suggests that maternal drug abuse before or during pregnancy is associated with more than twice the risk of fractures due to assault and transport accidents, compared with no maternal drug abuse. Associations were more prominent for assault-related fractures before 6 months of age and transport-related fractures after 6 years. Compared with no drugs, all types of drugs were associated with an increased risk of fractures due to assault and transport accidents. This study raises the possibility that illicit drug use early in the lives of women may predict fractures in offspring.

The association between maternal drug use and risk of pediatric fracture has not been previously investigated. The few existing studies pertain to alcohol and are hypothesis generating.^9,10 A U.S. survey of 12,360 single-family households reported that maternal alcohol abuse was associated with 2.1 times the risk of serious injury and 2.8 times the risk of fracture in children under 18 years, compared with non-drinkers.⁹ In a nested case–control study of 2,456 children with long-bone fractures and 23,661 controls, a history of maternal alcohol misuse was associated with 2.3 times the odds of long-bone fracture before 5 years of age.¹⁰ Neither study ascertained the timing of alcohol use or considered illicit drug use, despite rising rates in pregnant women.^2,12 In our data, drug abuse combined with alcohol was associated with assault and transport-related fractures relative to no drug or alcohol abuse, but alcohol abuse alone was not.

Our study provides new evidence that maternal drug abuse may be associated with the risk of fractures in children, especially assault-related fractures. Although no study has yet examined the relationship between illicit drug use and fractures due to assault, some data suggest that maternal substance use is associated with child maltreatment.^21–24 In a study of 8,472 respondents to a mental health survey in Canada, maternal substance use was associated with more than 4 times the odds of severe physical abuse.²¹ A matched-cohort study of 357 Australian infants with a median follow-up of 4 years found that children exposed to maternal use of opiates, amphetamines, or methadone had 13 times the risk of harm, an association that was driven by physical abuse mainly.²² Some matched cohort studies suggest that maternal cocaine use during pregnancy is associated with 5 to 7 times the risk of child maltreatment the first 2 years of life.^23,24 In our data, the association between maternal drug abuse and assault-related fractures was stronger early in infancy, especially before 6 months of age. This finding is consistent with evidence that most children who sustain fractures from abuse are under 18 months.²⁵

Substance abuse was also associated with an increased risk of fracture due to transport accidents. Illicit drug use may be hard to detect in motor vehicle accidents, but our data suggest that drug abuse before or during pregnancy could be a marker of future risk. Drug use may hamper child safety practices during driving.^26,27 Data suggest that children involved in motor vehicle accidents are more likely to be unrestrained and seated in the front seat when drivers use illicit drugs or alcohol.^26,27 A recent study of child passengers involved in fatal motor vehicle accidents reported that children were 1.5 times more likely to be seated in the front seat if the driver used drugs.²⁷ Children were more likely to be unrestrained if the driver used cannabis.²⁷ These findings are concerning as several studies show that cannabis consumption increases the risk of motor vehicle crashes.²⁸ In our study, cannabis was strongly associated with transport-related fractures. Moreover, the risks strengthened with the age of the child.

Biological mechanisms may also be involved. Studies in animals and humans suggest that maternal drug use may adversely affect the developing fetal skeleton.^29–31 In rats, prenatal cocaine exposure results in reduced femoral weight and density in offspring, suggesting an effect of cocaine on bone cell function and mineral metabolism.²⁹ In humans, serum levels of osteocalcin, a marker of bone formation and remodeling, are lower in neonates of mothers who use heroin or cocaine, compared with no drugs.³⁰ Appetite suppressant or vasoconstrictive effects of drug use may interfere with fetal supply of nutrients required for bone mineralization, including vitamin D and calcium.^29,31 A UK study of 198 children found that antenatal vitamin D deficiency and decreased placental calcium transfer were associated with reduced bone mass several years after birth,³² potentially increasing the risk of fracture.³³ However, these pathways may not be as important as parenting style since we found strong associations with fractures due to assault and transport accidents. These fractures may be influenced by behavioral factors that impede parents from providing nurturing environments and mother–child attachments.^4,8,21–24

This study has limitations that are worth considering. As we analyzed hospital data, we could not account for women who used illicit drugs before pregnancy but did not develop a disorder requiring admission. Maternal drug abuse, including polysubstance abuse, may be underreported or misclassified. Misclassification of exposure may underestimate the association with childhood fractures. We did not have information on the severity of maternal drug abuse, including dose, frequency, and whether women continued, ceased, or started using drugs after delivery. Maternal drug use behavior may change after pregnancy. We could not determine whether women abused prescription drugs or were treated for substance use disorders. We focused on complicated fractures that required hospital admission but had no information on the circumstances surrounding the injury, including safety of the physical environment and level of supervision. Associations with assault-related fractures may be misestimated as child maltreatment is difficult to capture and potentially underreported.³⁴ While we accounted for several patient characteristics, we cannot rule out residual confounding as information on ethnicity, social support levels, single motherhood, and paternal drug use was not available in the data. The measure of socioeconomic status may not have captured the full range of adverse experiences due to deprivation. The findings of this study are representative of a Canadian pediatric population, but generalizability to settings with different demographics is unclear.

Conclusions

In this large cohort of Canadian children, maternal drug abuse before or during pregnancy was a predictor of future risk of hospitalization for fractures due to assault and transport accidents in offspring. The data indicate that despite hospitalization for substance abuse, risks to the child persisted. The findings suggest that despite early detection of substance abuse, current health services are not sufficiently addressing risks to the next generation. Greater efforts to screen and refer young women for treatment, education, and counseling in the healthcare system may be indicated. A stronger gender focus may be merited. A substantial number of women initiate substance use long before pregnancy, especially in adolescence,^11,35 and unplanned pregnancy may be common. The opioid crisis and liberalization of cannabis may lead to more substance use in women of childbearing age,^3,36 reinforcing the public health significance of our findings. Future research may benefit from more accurately capturing the burden of maternal substance use on childhood fractures.