Pregnancy, fetal, and neonatal outcomes among women with traumatic brain injury

Abstract

Objective:

There have been no systematic studies of pregnancy outcomes among women with traumatic brain injury (TBI), potentially limiting informed clinical care for women with such injuries. The purpose of this exploratory study was to evaluate pregnancy and fetal/neonatal outcomes among women with a TBI diagnosis recorded during their delivery hospitalization, compared to women without TBI.

Setting:

In this cross-sectional study, we identified women with delivery hospitalizations using 2004–2014 data from the Nationwide Inpatient Sample of the Health Care and Cost Utilization Project.

Participants:

We identified deliveries to women with a TBI diagnosis on hospital discharge records, which included all diagnoses recorded during the delivery, and compared them to deliveries of women without a TBI diagnosis.

Main Measures:

Pregnancy outcomes included: gestational diabetes; preeclampsia/eclampsia; placental abruption; cesarean delivery; and, others. Fetal/neonatal outcomes included: preterm birth; stillbirth; and, small or large gestational age.

Design:

We modelled risk for each outcome among deliveries to women with TBI, compared to women without TBI, using multivariate Poisson regression. Models included sociodemographic and hospital characteristics; secondary models added clinical characteristics (e.g., psychiatric disorders) which may be influenced by TBI.

Results:

We identified 3,597 deliveries to women with a TBI diagnosis and 9,106,312 deliveries to women without TBI. Women with TBI were at increased risk for placental abruption RR 2.73 (95% CI 2.26, 3.30) and associated sequelae (i.e., antepartum hemorrhage, cesarean delivery). Women with TBI were at increased risk for stillbirth RR 2.55 (95% CI 1.97, 3.29) and having a baby large for gestational age RR 1.30 (95% CI 1.09, 1.56). Findings persisted after controlling for clinical characteristics.

Conclusions:

Risk for adverse pregnancy outcomes, including placental abruption and stillbirth, were increased among women with TBI. Future research is needed to examine the association between TBI and pregnancy outcomes using longitudinal and prospective data, and to investigate potential mechanisms that may heighten risk for adverse outcomes.

INTRODUCTION

Traumatic brain injury (TBI) is a major cause of morbidity and mortality in the United States, contributing to approximately 2.9 million TBI-related emergency department visits, 288,000 hospitalizations, and 56,000 deaths in 2014.¹ These estimates are conservative since these numbers do not capture TBIs identified in outpatient settings, and many who are injured do not seek medical attention.² While men have higher age-adjusted TBI-related emergency department visits and hospitalization rates than women, rates are increasing for both men and women.¹ Mechanism of injury varies by gender, with women significantly more likely to experience TBI resulting from unintentional falls, intimate partner violence (IPV), and occupational injuries (e.g., health and social services).^1,3,4

There is limited research examining reproductive health outcomes or gender or sex-specific outcomes following TBI; with the few studies in this area limited by small sample sizes and restricted to those receiving care in rehabilitation settings.^5–9 Nevertheless, findings from these studies indicate that TBI may have unique biological and psychosocial consequences for women including greater risk for sexual dysfunction¹⁰ and amenorrhea and irregular menstrual cycles, possibly due to sustained damage to the pituitary gland,^6,7,11–13 and the hypothalamic-pituitary-gonadal axis.¹⁴ A recent study of women seeking care in the emergency department for a concussion, compared to those seeking care for extremity injuries, found that pregnancy incidence 24 months post-injury was 76% lower for women with a concussion compared to those without.¹⁴ Women with TBI have also reported reduced utilization of obstetrician/gynecologist services and family physician/midwife services post-injury, compared to pre-injury, which may have consequences for their reproductive health.⁵ Yet, to date, no studies have looked at pregnancy, fetal, or neonatal outcomes among women with TBI.

TBI may have acute implications for pregnancy outcomes related to the mechanism of injury. For example, IPV is associated with increased risk of placental abruption, antepartum hemorrhage, and stillbirth.¹⁵ Further, emerging evidence suggests that for some individuals with TBI, such injuries are better characterized as chronic health conditions that have life-long effects.¹⁶ That is, TBI-related sequelae including physical, cognitive, emotional, and behavioral impairments can be exacerbated by co-occurring conditions and barriers to TBI treatment stemming from social determinants of health (e.g., persons with lower incomes, racial/ethnic minorities, or those without health insurance),^17,18 as well as factors associated with normal aging.¹⁹ Acute and persistent TBI sequelae, even when the injury is mild, may trigger progressive processes (e.g., neurologic, inflammatory), resulting in increased risk for dementia, Parkinsonism, cognitive deficits, depression, and psychosis.²⁰ Some TBI-related sequalae, such as depression, are also potential risk factors for adverse pregnancy outcomes, including gestational diabetes, preeclampsia, low birthweight, and stillbirth.²¹ In addition to medical complications following TBI, persons with TBI may experience barriers to healthcare or reduced quality of care due to communication challenges which may be detrimental to pregnancy outcomes.

To date, there has been no systematic study of the association of TBI with pregnancy outcomes. Thus, the purpose of this exploratory study is to begin the process of closing this gap in knowledge by examining the associations of TBI with pregnancy and fetal/neonatal outcomes using data from the Nationwide Inpatient Sample of the Health Care and Cost Utilization Project (HCUP-NIS).

STUDY DATA AND METHODS

We conducted a retrospective analysis using 2004–2014 HCUP-NIS data, which is sponsored by the Agency for Healthcare Research and Quality and is the largest publicly available all-payer, inpatient health care database in the United States. HCUP-NIS contains data on approximately 8 million hospital stays per year from over 1,000 hospitals, yielding a stratified sample of approximately 20% of United States community hospitals. HCUP-NIS has been used to study pregnancy outcomes among sub-populations of women within the disability community.^22,23 This study used existing, de-identified secondary data, and was granted a waiver of approval from the Brandeis University Institutional Review Board.

The combined 2004–2014 HCUP-NIS contained 85,399,326 hospitalization records (see Figure 1). Of these, we identified a cohort of 9,109,909 delivery hospitalizations using Kuklina et. al’s²⁴ enhanced delivery identification method based on ICD-9-CM, shown to identify 3.4% more deliveries compared to other methods, including those with severe obstetric complications. From this cohort, we identified 3,597 deliveries to women with a TBI diagnosis using ICD-9 codes that were captured in the delivery discharge data (see Table I); deliveries to other women were used as the comparison group (n=9,106,312). Similar to other studies based on HCUP-NIS data examining comorbidities among women hospitalized for a delivery,²⁵ ICD-9-CM diagnosis codes recorded during the delivery and present on the discharge records were used to examine other comorbidities.

Figure 1.

This is a flow chart depicting the analytic sample selection from the total number of hospitalizations available in the 2004–2014 Nationwide Inpatient Sample of the Health Care and Cost Utilization Project data to the number of delivery hospitalizations included in our analyses. TBI = traumatic brain injury.

Table 1.

Definitions for traumatic brain injury, pregnancy and neonatal outcomes, and clinical characteristics

Variables	ICD-9 diagnosis and procedure codes
Traumatic Brain Injury	310.2; 800.xx; 801.xx; 803.xx; 804.xx; 850.xx-854.xx; 950.1x-950.3x; 959.01; V15.5; V15.59; V15.52
Pregnancy complications
Gestational diabetes	648.8
Preeclampsia and eclampsia	642.4, 642.5, 642.9
Placenta previa	641.0, 641.1, 762.0
Placental abruption	641.2, 762.1
Labor and Delivery Outcomes
Breech	652.2, 669.6, 763.0
Cesarean delivery	669.7, 763.4, 74, 74.1, 74.2, 74.4, 74.9
Antepartum hemorrhage	641.1× 641.2× 641.3× 641.8× 641.9x
Postpartum hemorrhage	666.1x; 666.0× 666.2× 666.3x
Fetal and neonatal outcomes
Small for gestational age/Poor Fetal growth	656.5x
Large for gestational age/Excessive fetal growth	656.6x
Preterm birth (<37 weeks)	644.2, 644.20, 644.21, 765.0, 765.1
Having a stillbirth	768.0, 768.1, 656.4x V27.1, V27.3, V27.4, V27.6, V27.7
Clinical Characteristics
Substance use disorder	291.xx 292.xx 303.xx-305.xx 648.3× 655.5× 965.0x V65.42
Psychiatric Conditions
Anxiety Disorder	300.0x; 300.2x; 300.3x; 309.20; 309.21; 309.24; 309.81
Bipolar Disorder	296.0x; 296.1x; 296.4x; 296.5x; 296.6x; 296.7x; 296.80; 296.81; 296.89; 296.9x; 301.13
Attention Deficit Disorders	314.xx
Dementia	290.xx; 780.93x
Depression	296.2x;296.3x;296.82;298.0x;300.4x;301.12; 309.0x;309.1x;309.28; 311.xx
Obsessive Compulsory Disorder	300.3x
Other psychoses	297.1x; 297.3x; 298.8x; 298.9x; 301.22
Schizophrenic disorders	295.xx
Suicide attempts E950 -E958	E950 -E958

Source: Healthcare Cost and Utilization Project National Inpatient Sample (HCUP-NIS) data, 2004–2014

Abbreviations: ICD-9 International Classification of Diseases and Related Health Problems 9th Revision.

Independent Variable

The independent variable was presence of a TBI diagnosis on the delivery hospitalization discharge record identified by use of ICD-9-CM diagnosis codes.

Outcomes

The study outcomes included pregnancy complications, outcomes of labor and delivery, and fetal, and neonatal outcomes identified by ICD-9-CM codes (Table 1).^26,27 Pregnancy complications included gestational diabetes mellitus; preeclampsia and eclampsia; placenta previa; and, placental abruption. Outcomes of labor and delivery included: breech; cesarean delivery; antepartum hemorrhage; and, postpartum hemorrhage. Fetal and neonatal outcomes included: preterm birth (<37 weeks); stillbirth; small for gestational age; and, large for gestational age. Outcomes were measured as dichotomous variables with 1 indicating an adverse outcome.

Covariates

Covariates were selected based on prior research and included sociodemographic, hospital, and clinical characteristics associated with TBI and/or pregnancy, fetal, and neonatal outcomes.^28–30 Sociodemographic characteristics included: maternal age (≤19, 20–34, 35+); race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic, Non-Hispanic Other, Unknown race/ethnicity); health insurance (Medicare, Medicaid, private, uninsured); and, median household income for patients’ zip code (0 to 25th percentile, 26th to 50th percentile, 51st to 75th percentile, 76th to 100th percentile). Hospital characteristics included hospital capacity (i.e., number of short-term acute beds; small: 1–49 beds; medium: 50–99 beds; large: ≥100 beds); and geographic location (Northeast, Midwest, South, West). Clinical characteristics included: substance use disorder (yes/no); psychiatric disorder (yes/no); and, Elixhauser comorbidity score based on Agency for Health Care Research and Quality Elixhauser Comorbidity Index.³¹ The Elixhauser comorbidities are 29 clinical conditions (e.g. diabetes, obesity, hypertension, etc.) that exist prior to a patient’s admission to the hospital and are not related to the principal reason for hospital admission, but may be significant factors influencing hospitalization outcomes.³¹ We categorized the Elixhauser Comorbidity Index and measured it as a categorical variable (none, 1, 2, 3 or 4+). Year of delivery was included in the models.

Statistical Analysis

We examined differences in sociodemographic, hospital, and clinical characteristics between the cohort of deliveries to women with and without a TBI diagnosis. For categorical valued characteristics, weighted frequencies and proportions were reported. For continuous characteristics, weighted means and standard deviations were reported.

To evaluate the association between TBI status and the risk of adverse outcomes, we used Poisson regressions for binary outcomes and estimated unadjusted and adjusted relative risks (RR) and 95% confidence intervals (CI). To avoid over-adjustment bias, the main models adjusted for sociodemographic and hospital characteristics, and year of delivery (Model 1). The second set of multivariable models (Model 2), added clinical characteristics to assess how they influence the results from Model 1.

Analyses adhered to the methodological standards for research using the HCUP-NIS.³² HCUP-NIS hospital discharge weights were applied to the sample data for descriptive statistics to create national estimates.³³ Due to changes in the HCUP-NIS sampling design in 2012, we adjusted hospital discharge weights from prior years using the new hospital discharge weights.³⁴ Less than 1% of the sample was missing data for both health insurance and hospital bed size, and less than 2% for median household income. To address potential bias due to missing covariate data we used multiple imputation by chained equations to impute values for the variables with missing data.³⁵ The multiple imputations chained equation included a multiple logit model for insurance as the outcome and ordered logit models for income and hospital bed size as outcomes. Covariates for all these models included TBI status and all covariates with non-missing values, including age, race, region of the hospital, and number of comorbidities. We added 5 imputations for each missing value. We conducted sensitivity analyses with cases excluded where there was missing covariate data to determine if our findings were impacted by using multiple imputation. To address missing values for race/ethnicity (17%), we created an unknown race/ethnicity category. All models were estimated using Stata version 16 MP.

RESULTS

Bivariate analyses revealed that women with a TBI diagnosis recorded during their delivery hospitalization were more likely to be non-Hispanic white and to have Medicare health insurance, compared to women without TBI (Table II). Women with TBI were more likely to have psychiatric and substance use disorders, and a greater number of Elixhauser comorbidities than those without TBI.

Table II.

Sample Characteristics Among Deliveries to Women with TBI and Women in General Obstetric Population, United States, 2004–2014

Characteristics	Women with TBI N=3,597		Women without TBI N=9,106,312
	Unweighted N	Weighted %	Unweighted N	Weighted %
Sociodemographic
Maternal age at delivery
<20	299	8.3	842,247	9.2
20–24	852	23.7	2,171,187	23.8
35–39	2,251	62.6	5,808,619	63.9
40+	195	5.4	284,259	3.1
Maternal age at delivery, (Mean, SD)	27.4	1.0	27.8	0.1
Race and ethnic identity
Non-Hispanic White	2,144	59.7	3,967,880	43.7
Non-Hispanic Black	331	9.2	1,033,754	11.3
Hispanic	470	13.1	1,753,597	19.2
Non-Hispanic Other	215	6	814,831	9
Unknown Race/Ethnicity	437	12.1	1,536,250	16.9
Health insurance
Medicare	159	4.5	62,687	0.7
Medicaid	1,551	43.2	3,868,903	42.4
Private	1,684	46.7	4,605,434	50.6
Uninsured	198	5.4	553,534	6.1
Missing	5	0.2	15,754	0.2
Median household income for patient’s Zip code
0 to 25th percentile	966	26.8	2,445,103	26.9
26th to 50th percentile	955	26.5	2,256,390	24.8
51st to 75th percentile	890	24.8	2,187,525	24.0
76th to 100th percentile	723	20.1	2,046,977	22.5
Missing	63	1.8	170,317	1.9
Clinical characteristics
Number of Elixhauser comorbiditiesa
None	2,034	56.3	6,987,757	76.6
1	836	23.3	1,242,414	13.7
2	454	12.7	664,638	7.3
3	181	5	158,974	1.8
4+	92	2.6	52,529	0.6
Psychiatric disorders	546	15.3	383,358	4.2
Substance use disorder	244	6.8	230,242	2.5
Hospital characteristics
Hospital bed size
Small	452	12.2	1,113,120	11.7
Medium	810	22.6	2,433,689	26.9
Large	2,307	64.5	5,513,178	60.9
Missing	28	0.8	43,625	0.5
Region of hospital
Northeast	475	13.5	1,468,888	16.4
Midwest	815	22.9	1,903,750	21.3
South	1,402	38.5	3,479,064	37.8
West	905	25.2	2,254,610	24.6

Source: Healthcare Cost and Utilization Project National Inpatient Sample (HCUP-NIS) data, 2004–2014

Abbreviations: TBI-Traumatic Brain Injury.

Notes: Data unweighted n (weighted %) unless otherwise specified.

^aNumber of Elixhauser comorbidities was based on the Agency for Health Care Research and Quality (AHRQ) standard methods by Elixhauser.

In the main models (Table III), women with TBI had increased risk for placental abruption RR 2.73 (95% CI 2.26, 3.30), antepartum hemorrhage RR 2.32, (95% CI 1.95, 2.75), and cesarean delivery RR 1.83 (95% CI 1.75, 1.91). These reproductive health outcomes persisted after controlling for clinical characteristics. In the main models (Table IV), women with TBI were at increased risk for having a stillbirth RR 2.55 (95% CI 1.97, 3.29), and having a neonate large for gestational age RR 1.30 (95% CI 1.09, 1.56) compared to women without TBI. These associations persisted after controlling for clinical characteristics.

Table III.

Pregnancy Outcomes among Deliveries to Women with and without Traumatic Brain Injury, 2004–2014

Outcomes	Women with TBI		Women without TBI		Model 1a		Model 2b
	Unweighted N	Weighted %	Unweighted N	Weighted %	RR	95% CI	RR	95% CI
Pregnancy Complications
Gestational diabetes mellitus	202	5.6	520,663	5.7	0.91	(0.79, 1.04)	0.88	(0.77, 1.01)
Preeclampsia and eclampsia	137	3.8	376,753	4.1	0.88	(0.75, 1.05)	0.78	(0.66, 0.92)
Placenta previa	16	0.4	49,401	0.5	0.79	(0.49, 1.30)	0.71	(0.43, 1.16)
Placental abruption	106	2.9	96,628	1.1	2.73	(2.26, 3.30)	2.28	(1.89, 2.76)
Labor and Delivery Outcomes
Breech	107	3.0	277,853	3.1	0.97	(0.81, 1.18)	0.93	(0.77, 1.12)
Cesarean delivery	2,155	59.9	2,939,091	32.3	1.83	(1.75, 1.91)	1.74	(1.67, 1.82)
Antepartum Hemorrhage	131	3.6	141,174	1.6	2.32	(1.95, 2.75)	1.97	(1.66, 2.33)
Postpartum Hemorrhage	103	2.9	262,598	2.9	0.98	(0.81, 1.19)	0.84	(0.70, 1.02)

Source: Healthcare Cost and Utilization Project National Inpatient Sample (HCUP-NIS) data, 2004–2014

Abbreviations: TBI-Traumatic Brain Injury; CI=Confidence Intervals; RR=Risk Ratios

Notes: Data unweighted n (weighted %) unless otherwise specified.

^aAdjusted for maternal age, race/ethnicity, insurance, median income for patient’s zip code, hospital bed size, region, and birth year;

^bAdjusted for covariates in the Model 1 plus Elixhauser comorbidity score, substance use disorder, and psychiatric disorders.

Table IV.

Fetal and Neonatal Outcomes among Deliveries to Women with and without Traumatic Brain Injury, 2004–2014

Outcomes	Women with TBI		Women without TBI		Model 1a		Model 2b
	Unweighted N	Weighted %	Unweighted N	Weighted %	RR	95% CI	RR	95% CI
Preterm birth	279	7.8	654,238	7.2	1.10	(0.98, 1.24)	0.98	(0.88, 1.11)
Having a stillbirth	58	1.6	56,039	0.6	2.55	(1.97, 3.29)	2.29	(1.77, 2.97)
Small for gestational age	85	2.4	196,155	2.2	0.99	(0.80, 1.23)	0.90	(0.73, 1.11)
Large for gestational age	119	3.3	234,312	2.6	1.30	(1.09, 1.56)	1.26	(1.05, 1.51)

Source: Healthcare Cost and Utilization Project National Inpatient Sample (HCUP-NIS) data, 2004–2014

Abbreviations: TBI-Traumatic Brain Injury; CI=Confidence Intervals; RR=Risk Ratios

Notes: Data unweighted n (weighted %) unless otherwise specified.

^aAdjusted for maternal age, race/ethnicity, insurance, median income for patient’s zip code, hospital bed size, region, and birth year;

^bAdjusted for covariates in the Model 1 plus Elixhauser comorbidity score, substance use disorder, and psychiatric disorders.

Findings remained unchanged in sensitivity analyses that excluded cases where multiple imputation was used to address missing data (see Supplemental Digital Content Tables 1 and 2).

DISCUSSION

Women with a TBI diagnosis recorded on their delivery hospital discharge were at increased risk for placental abruption and associated sequelae (i.e., antepartum hemorrhage, cesarean delivery), compared to women without a TBI after controlling for sociodemographic and hospital characteristics. Additionally, TBI was associated with increased risk of stillbirth and having a neonate large for gestational age. These findings persisted after also controlling for clinical characteristics.

Prior literature has found that physical trauma, such as IPV or motor vehicle accidents, increases risk of placental abruption and associated sequalae,^36,37 as well as perinatal death or stillbirth.³⁸ Notably, these types of physical trauma, which are also common mechanisms of injury for TBI,^1,3,4 may be acute (e.g. motor vehicle accidents) or on-going (e.g., IPV), with potential short and long-term health and behavioral consequences. While HCUP-NIS data does not include information on the mechanism or severity of injury, or the timing of TBI in relation to the delivery, we hypothesize that the TBI diagnoses recorded on the hospital delivery discharge data were important enough that they were communicated to the clinical labor and delivery team. Thus, they likely capture more acute injuries that occurred during or close to the pregnancy, or more severe injuries that occurred prior to the pregnancy and continued to present disabling effects that are communicated with all providers. We can only speculate as to whether the observed associations are a direct effect of TBI as a chronic condition, or reflective of risks associated with the injury that led to TBI. Also of note, ongoing work highlights multiple pathways associated with placenta abruptions including chronic inflammation and immunodysregulation,³⁹ processes also implicated in negative outcomes post-TBI.⁴⁰

The association of TBI with increased risk of having a neonate large for gestational age was unexpected and suggests that TBI may have impacts on neonatal morbidity and mortality separate from placental abruption, which is typically associated with growth restriction.^41,42 Women with a TBI diagnosis were more likely that those without TBI to have pre-existing medical conditions (i.e., psychiatric and substance use disorders, Elixhauser comorbidities). This is consistent with well-documented literature demonstrating that persons with TBI are at increased risk for post-injury substance use and mental health problems, and poorer physical health.^43,44 Some of these comorbidities, particularly diabetes are recognized risk factors for having neonates large for gestational age. Women with a TBI diagnosis in our study had an almost 60% higher prevalence of pre-existing diabetes diagnoses on their hospital discharge records, compared to women without TBI (see Supplemental Digital Content Table 3). However, if the association of TBI and increased risk for having a neonate large for gestational age was driven by increased likelihood of women with TBI having diabetes and related risk factors, we would have expected to also observe an association between TBI and gestational diabetes. However, such an association was not observed.

One of the most striking results was the over two-fold increased risk of stillbirth associated with TBI. Stillbirth is a rare but devastating outcome of pregnancy that remains poorly understood and occurs in approximately 1 of 160 deliveries in the United States.⁴⁵ Many of the physical and mental health comorbidities associated with TBI (e.g., depression, substance use) are also risk factors for stillbirth,⁴⁵ yet these findings persisted once clinical characteristics were included in the models. This also may be a result of the observed association between TBI and placental abruption given the high correlation between severity of placental abruption and stillbirth.^46,47 Alternatively, social determinants of health (e.g., poor nutrition, experiences of violence, high exposure to air pollution) are also known to increase risk of stillbirth and may partially explain the observed association if TBI was a marker for some of these same risks (e.g., IPV).⁴⁸ Nevertheless, as we cannot determine the mechanism of injury in these data, more research is needed to investigate this consequential finding.

HCUP-NIS provides high-quality, nationally representative data and permits us to draw inferences about the population of women with TBI in the United States who had a delivery hospitalization. Findings are at the delivery-level, not the person-level, therefore it is possible that some women may contribute to multiple deliveries during the study period. The sample is less constrained by selection bias or sampling bias common in convenience samples derived from a single hospital, service provider organization, geographic region, or rehabilitation-seeking population. Yet, there are limitations that warrant consideration. As reported in other studies that use HCUP-NIS data to examine comorbidities and diagnoses among women hospitalized for deliveries,²⁵ claims that do not generate a specific payment, such as ICD-9-CM codes for TBI or chronic conditions, may be under-recorded. Thus, our estimates are likely to be conservative. As discussed above, we do not know the timing of when the TBI occurred, including whether it occurred during pregnancy, in the prenatal phase, or earlier in life. We did confirm that only 0.4% of deliveries to women with a TBI diagnosis had the TBI diagnosis recorded as the primary diagnosis on the hospital discharge record (data not shown); thus, it was very uncommon for the TBI to be the reason the woman was seeking care in the hospital when the delivery occurred. Similar to other TBI research of adults in the United States that relies on electronic medical record data to capture TBI diagnoses, we speculate that the estimates for lifetime history of TBI are conservatively biased because TBI is not always recorded in the HCUP-NIS record, particularly if women did not seek care related to their TBI or discuss their injury history with their clinicians during pregnancy. This may be more likely among women with a history of mild TBI. As such, the indicator for no TBI may misclassify some women who do have a history of TBI. This misclassification would likely reduce any differences (shift towards the null) between women with TBI and those without TBI. Additionally, we do not know mechanism of injury, nor whether the etiology of injury that resulted in the TBI also directly increased risk for adverse pregnancy outcomes (e.g., placental abruption) rather than the TBI itself increasing this risk. Due to restrictions related to available HCUP-NIS data, the team was unable to account for marital status, a variable that has unique impacts on birth outcomes.⁴⁹ Lastly, since the data is cross-sectional, causality cannot be asserted regarding relationships found in this study, only associations.

Our findings add to emerging studies suggesting that women with TBI may experience impairments to their reproductive health, including the recent study, Pregnancy After Concussion: A Clarion Call for Attention?,¹⁴ which found that women with a concussion had a reduced incidence of pregnancy compared to women without concussion. Taken together, our new findings highlight the need for future research to inform increased understanding regarding pre-existing risk, and associated negative reproductive health outcomes among women with TBI using longitudinal and prospective epidemiological and biological data. This will require adoption of paradigms which not only appreciate risk during the time of pregnancy, but also incorporate the accumulation of risk over the lifetime.⁵⁰ For example, a recent population-based cohort study using administrative healthcare data from 2003–2018 in Ontario, Canada found that women with physical disabilities, inclusive of brain injury, had an increased rate of severe maternal morbidity (i.e., a validated composite measure of 40 diagnostic and procedural indicators) as well as all-cause maternal mortality, compared to women without a disability.⁵¹ Another study⁹ found that among a sample of women with moderate-severe TBI, there was a trend towards increased mortality among women who were pregnant. More research is needed with larger samples to investigate risk for this devastating outcome. Future research is needed to examine these outcomes and other postpartum health outcomes by TBI status, or by mechanism of injury (e.g., IPV, vehicle accident). Modeling of accumulated risk will require increased understanding regarding sex-effects (e.g., potential differences in inflammation) on TBI outcomes (e.g., onset or exacerbation of psychiatric conditions post-injury) and the impact of these on reproductive health and outcomes across the life-course.⁵² Towards this end, evidence is mounting that suggests that screening for lifetime history of TBI should be considered a standard primary and reproductive healthcare procedure for OB/GYNs and midwives, much like screening for depression or substance use.

Moreover, additional research is needed to explore the impact of TBI severity on reproductive health outcomes, timing of TBI in relation to pregnancy, whether women with a history of TBI are at increased risk for experiencing IPV before or during pregnancy, and to understand the impact of social determinants of health on pregnancy and birth outcomes among women with TBI. Additional study is warranted to follow infants beyond data collected in HCUP-NIS, to explore if infants of mothers with TBI have increased risk for adverse developmental or long-term health outcomes (e.g., childhood obesity, mental health conditions). A recent pre-clinical study found that offspring of pregnant mice injured with a TBI during pregnancy had increased anxiety- and depression-related behaviors and altered immune responses compared to offspring of non-injured mice, and male offspring had worse outcomes.⁵³ The authors hypothesized that peripheral inflammation and other psychopathology of the TBI-injured, pregnant mother can travel through the blood-placenta barrier to impact organ systems throughout the developing fetus.⁵³ Future research based on detailed prenatal care data could more precisely explore the relationships between early prenatal care and pregnancy outcomes.

Additionally, future studies should include primary data collection to understand if women with TBI experience unique barriers to prenatal care, communication challenges during visits, or stigma, that may impact the quality of prenatal care received. Lastly, more research is warranted to understand the perinatal experiences of women with TBI, which may reveal barriers to care and identify areas for improvement.

CONCLUSIONS

Findings from this study continue the important, yet nascent process of exploring the association of TBI on reproductive health outcomes. Research that informs our understanding of women’s reproductive health and pregnancy outcomes following TBI is expected to provide rehabilitation and reproductive health providers with information to improve education for clinicians regarding potential pregnancy complications for women with TBI. Women with a TBI diagnosis recorded during their delivery hospitalization were more likely to have both pre-existing comorbidities, as well as adverse pregnancy and fetal/neonatal outcomes. Results support the need for development of future longitudinal and prospective studies to understand the association of history of TBI with pregnancy outcomes, where timing of TBI in relation to pregnancy and mechanism of injury are more clearly delineated. If future research replicates and extends our finding of high-risk pregnancy outcomes for women with TBI, women with a history of TBI may benefit from closer monitoring to optimize maternal and fetal outcomes, and provider education will be warranted for those working in TBI rehabilitation. Routine screening for TBI during pre-conception and early in pregnancy may help identify women at higher risk for placental abruption and stillbirth.