History of Strangulation Is Associated with Current Traumatic Stress, Self-Reported Vision Problems, and Other Neurobehavioral Symptoms in Women Who Have Experienced Intimate Partner Violence

Abstract

Intimate partner violence (IPV) is a serious and prevalent problem affecting approximately one in three women globally. Physical IPV can involve non-fatal strangulation (NFS), which can result in an acquired brain injury (ABI), inferred by an alteration in consciousness (AIC). However, there is limited research pertaining to NFS-related ABIs, especially in the context of understanding long-term outcomes. We examined neurobehavioral and traumatic stress symptoms associated with a past history of IPV-related strangulation, focusing on the presence of strangulation and the presence and type of a strangulation-related AIC. A sample of 139 women aged 18 years and older (mean = 40 years) was recruited via flyers shared with community partners (e.g., domestic violence advocates) as well as online advertisements and social media. Assessments included the Brain Injury Severity Assessment, Ohio State University Traumatic Brain Injury Identification Method, revised Conflict Tactics Scale, Neurobehavioral Symptom Inventory, and Post-traumatic Stress Disorder (PTSD) Checklist for the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5), the PCL-5. Multivariable linear and ordinal logistic regression models were used to examine the association between strangulation history and neurobehavioral and traumatic stress symptoms. The average time since the most recent strangulation was 8.7 years. Women who experienced strangulation reported greater levels of self-reported vision problems (p < 0.01) and traumatic stress symptoms (p < 0.05) than women who were not strangled, although only vision problems remained significant (p < 0.01) after adjusting for confounders. Women who sustained an AIC during strangulation reported higher levels of current traumatic stress symptoms, dizziness, vision problems, and poor concentration (p < 0.05; p < 0.05; p < 0.01; p < 0.01, respectively) in comparison with women who were strangled but did not sustain a related AIC; after adjusting for confounders, vision problems remained significant (p < 0.01). Women who specifically sustained a loss of consciousness (LOC) compared to another type of AIC, reported higher levels of traumatic stress symptoms (p < 0.05), even after adjusting for potential confounders. We found that long after the most recent IPV-related strangulation event, the presence of strangulation, strangulation-related AIC, and strangulation-related LOC were associated with a range of neurobehavioral symptoms and traumatic stress symptoms. However, after adjusting for potential confounders, strangulation and strangulation-related AICs were associated with self-reported vision problems, and strangulation-related LOC was associated with traumatic stress. This study highlights the potential long-term consequences of IPV-related strangulation and reinforces the importance of IPV prevention and providing treatment for these women in need.

Introduction

Intimate partner violence (IPV), defined as aggressive and/or abusive behavior of a physical, sexual, or psychological nature in an intimate relationship,¹ is a persistent and serious problem worldwide. The World Health Organization estimates that nearly one in three women globally have reported experiencing IPV in their lifetime.² Many women who experience IPV sustain IPV-related acquired brain injuries (ABIs), which include traumatic brain injuries (TBIs; e.g., injuries to the head from a traumatic external force) and ABIs from non-fatal strangulation (NFS).³ Studies have revealed associations between these ABIs and negative psychological and cognitive outcomes for IPV survivors. The psychological outcomes associated with these ABIs include depression, anxiety, and traumatic stress.^4–8 In addition, ABIs have been shown to be negatively associated with scores on cognitive tests of memory, cognitive flexibility, learning, and executive/motor function.^9–11

NFS is a common form of IPV, defined as “the external compression of a person’s neck and/or upper torso in a manner that inhibits that person’s airway or the flow of blood into or out of the head.”¹² Wilbur and colleagues (2001) found that 68% of women who experienced IPV had been strangled by an intimate partner in their lifetime.¹³ Even though most cases of strangulation do not result in visible external injuries, some women who experience IPV are thought to incur strangulation-related brain injuries, with associated documentable neurobehavioral symptoms.¹⁴ Researchers have reported that strangulation in women who experience IPV is associated with traumatic stress, depression, suicidality, and lower scores on tests of memory compared with women who have never experienced strangulation.^15–17

Despite the previous work, most studies have focused on strangulation more broadly and not on whether there is also an associated alteration in consciousness (AIC).^{12–16,18–21} Suffering an AIC during NFS may result in hypoxic-ischemic brain injury if brain cells are deprived of oxygen and nutrients due to the compression of the carotids or trachea.^14,22 Partially because of its high metabolic demand, the brain is susceptible to damage caused by the deprivation of blood and/or oxygen supply.^23,24 Women who have suffered a strangulation-related AIC have been shown to have worse psychological and cognitive outcomes than women who have not experienced a strangulation-related AIC.^18,25,26 However, these studies are limited in number and scope of analysis, comparing solely women who had experienced strangulation-related AICs with women who had not or studying just the number of lifetime strangulation-related losses of consciousness (LOCs).²⁵ In addition, most IPV-related strangulation studies have been conducted in the acute, subacute, or post-acute time period after the event.^19,20,27,28

The current study aimed to provide further insights into strangulation-related AICs and LOCs experienced remotely in time as potential predictors of later neurobehavioral and traumatic stress symptoms in women who experienced IPV. As a result of our focus on IPV-related strangulation, we treated IPV-related TBIs caused by blunt force trauma as a covariate, adjusting for them in regression analyses. For this study, we chose to assess the subjective measures of functioning commonly considered in brain injury research as well as the outcomes that have been shown to be associated with IPV-related BIs more generally (i.e., TBIs and strangulation-related AICs). As such, the possible chronic outcomes we focused on included dizziness, vision problems, hearing difficulty, poor concentration, and feelings of anxiety and depression, as these neurobehavioral and mental health outcomes have been shown to be associated with IPV-related BIs.^6,8,29 We also examined traumatic stress, given its known association with IPV-BIs and strangulation.^8,26

The primary aims of this study were to analyze the aforementioned neurobehavioral symptoms and traumatic stress in women who experienced (i) IPV, comparing those with and without strangulation; (ii) IPV-related strangulation, comparing those with and without an AIC; and (iii) IPV-related strangulation with an AIC, comparing those with and without LOC during strangulation.

Methods

Sample

Women were recruited via flyers shared with community partners (e.g., domestic violence advocates), online advertisements (Mass General Brigham [MGB] Rally), and social media (e.g., Facebook/Instagram). MGB’s Rally website advertises research studies to all MGB staff—including physicians, researchers, clinical practitioners, and support staff—and the general public. We included all women aged 18 years and older who expressed interest, were capable of participating (e.g., not experiencing symptoms of psychosis, English-speaking), had appropriate technology to complete an online remote study (or could come to MGH to use laboratory technology), and had experienced at least one instance of physical partner violence. This study was conducted completely online to improve accessibility for study participants. This was particularly relevant during the beginning of the study when COVID-19 concerns made in-person visits either impossible or highly difficult. Data for this study were collected from July 2020 to December 2023 for 139 total study participants; 137 participants identified as cis-gender women and 2 identified as transgender women. There were a small number of women who reported having a lifetime history of a neurological condition or disorder, such as convulsions/epilepsy (n = 6), fainting spells (n = 4), stroke (n = 6), and confusional states (n = 2). All women were provided with remuneration in the form of a $50 e-Gift card. Due to the minimal renumeration amount, the length of the protocol, and the difficult and sensitive content matter, we do not believe that our study contains fraudulent participants. Nonetheless, we spot-checked the data throughout aspects of the protocol to ensure consistency, where possible. While doing this, we found that one woman who participated in a part of the protocol twice and had inconsistent data; her data were not included in any of the analyses. All women provided written informed consent by way of an e-signature.

Procedure

This study was approved by the MGB Human Research Committee. Women were invited to participate in an online study pertaining to the effects of partner violence on women’s health; the average study duration was 2.5–3.5 h. The study sessions included demographics interviews, cognitive testing, online questionnaires, and clinical interviews (including abuse history interviews) conducted by female research assistants. All assessors shadowed at least three assessments conducted by experienced test administrators and were assessed for proficiency by the study Principal Investigator (PI) before conducting any assessments on their own. Participants provided written informed consent through e-signature, with all the procedures and risks also explained verbally by the assessor. Measures were taken to accommodate participants within the online study environment, including frequent breaks and comprehensive support for digital application usage.

Measures

The Brain Injury Severity Assessment (BISA)⁸ is well-grounded in the literature of IPV-related brain injury^{8–10,17,30–32} and specifically designed for use with women who have experienced IPV and possible strangulation-related AICs. To identify AICs, it uses the definition developed by the American Congress of Rehabilitation Medicine Special Interest Group on Mild Traumatic Brain Injury (1993).³³ Specifically, the BISA assesses for “a traumatically induced physiological disruption of brain function, as manifested by at least one of the following: any loss of consciousness; any loss of memory for events immediately before or after the accident; any alteration in mental state at the time of the accident (e.g., feeling dazed, disoriented, or confused); focal neurological deficit(s) that may or may not be transient” (p. 86). In the case of strangulation, instead of the AIC occurring after external forces to the head, the AIC would need to occur after or during a strangulation incident where there is compression of the neck. The BISA records incidents of brain injury, including the severity of the incident (e.g., duration of LOC), how many incidents had occurred, and the recency of the incidents. Brain injuries were determined to be “mild” if the duration of LOC was ≤30 min and post-traumatic amnesia was ≤24 h. Notably, the BISA records strangulation-related AICs, indicating both the presence/absence of the incident and the total number of incidents. These AICs include LOC, dizziness, confusion/disorientation, seeing stars or spots, and/or memory loss surrounding the incident. In this study, we used AICs as a proxy for all ABIs, including strangulation-related ABIs. Consistent with its typical administration, the BISA was administered as an interview.

The Ohio State University Traumatic Brain Injury Identification Method (OSU-TBI-ID)³⁴ is a commonly used and well-validated structured interview designed for recording the lifetime history of TBI. It includes questions about specific possible brain injury mechanisms (e.g., falls, fights) and follow-up questions about the presence and duration of LOC, feeling dazed, or experiencing memory gaps following potential brain injury mechanisms. We combined the data collected from the BISA and OSU-TBI-ID to determine the most accurate estimate of IPV-related TBI and strangulation-related ABI (e.g., strangulation with an AIC including LOC). This was done by including all unique incidents indicated in either measure.

The revised Conflict Tactics Scale (CTS2)³⁵ is a validated 39-item assessment that measures the extent of physical, sexual, or psychological violence within an intimate relationship; it asks participants to fill in how many times their partner performed each act and whether it occurred in the past year. We used the CTS2 item “Your partner choked you” to provide information on whether a participant was ever strangled by an intimate partner, irrespective of whether an AIC co-occurred. Thus, comparison of strangulation-related AIC data from the BISA and strangulation data from this item allowed us to determine whether a survivor was strangled without suffering an AIC.

The Neurobehavioral Symptom Inventory (NSI)³⁶ was used to assess our a priori neurobehavioral symptoms of interest. The NSI is a validated 22-item self-report questionnaire that assesses a range of symptoms that the participant may have experienced over the past 2 weeks, additionally asking for the time of symptom onset. It uses a scale ranging from 0 to 4, with 0 indicating an intensity of “none” and 4 indicating the symptom as being “very severe.” From the questionnaire, we selected the following symptoms: dizziness (feeling dizzy), vision problems (vision problems, blurring, trouble seeing), hearing difficulty, poor concentration (poor concentration, can’t pay attention, easily distracted), feelings of anxiety (feeling anxious or tense), and feelings of depression (feeling depressed or sad). This measure has been used in previous studies of IPV.^31,37 During the study visit, the NSI was administered verbally, with the scale displayed to the subject via screen share, and if the subject endorsed a symptom, she was asked when the symptom began.

The Post-traumatic Stress Disorder (PTSD) Checklist for the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5), the PCL-5³⁸, is a validated 20-item self-report questionnaire that measures the DSM-5 symptoms of PTSD; each symptom has a 0–4 rating scale, with 0 indicating “Not at all” and 4 indicating “Extremely.” The PCL-5 includes a total symptom severity score ranging from 0 to 80. This measure has been used in previous studies of IPV.^17,26,30

The Alcohol Use Disorders Identification Test (AUDIT)³⁹ is a validated 10-item questionnaire that assesses alcohol consumption, drinking behaviors, and other alcohol-related problems within the past year. For this study, participants completed a self-report version. The AUDIT provides a score from 0 to 40, with 0 indicating a person who is abstinent from alcohol and has never had any alcohol-related problem within the past year.

Statistical analysis

The Wilcoxon rank sum test was used for continuous variables, and Fisher’s exact tests were used for categorical variables. Wilcoxon r and Cramer’s V were used for effect size measures. Spearman’s rank correlation was used to analyze the association between the number of AICs and the NSI and PCL-5 outcome variables, with the significance level of p ≤ 0.05. The effect size measures correspond as such: 0.1–0.3 = small effect, 0.3–0.5 = moderate effect, and ≥0.5 = large effect.

Logistic regression was used to identify significant associations between potential confounders (including age, years of education, traumatic stress symptomatology, history of partner-related TBI [presence/absence], alcohol intake in the past year, and presence of potentially confounding neurological disorders (e.g., convulsions/epilepsy, fainting spells, stroke, and confusional states), and our variables of interest. Any variables showing associations with p ≤ 0.10 were included in the relevant regression models. The presence of IPV-related TBI and a confounding neurological disorder was indicated as a zero or one depending on whether a TBI or neurological disorder was absent or present, respectively.

Multivariable linear and ordinal logistic regression (OLR) models were used for our primary objectives, namely examining between-group differences in neurobehavioral and traumatic symptoms for women who experienced the following: (i) strangulation versus no strangulation, (ii) strangulation with an AIC versus strangulation without an AIC, (iii) strangulation with LOC versus strangulation with an AIC other than LOC. Covariates were included as appropriate (as indicated above). Regression coefficients for the multivariable linear regression models and odds ratios for the multivariable OLR models are reported with 95% confidence intervals. When calculating the odds ratios for the multivariable OLR models, the reference for all binary variables (e.g., presence/absence, Y/N) was “No.”

For the OLR models, multicollinearity was assessed using variance inflation factors (with 4.0 as a threshold), and proportional odds assumptions were checked in SPSS using the Brant test. Assumptions were satisfied for all regressions. Constant variance was assessed for linear regression models using a fitted value versus a residuals plot. Because this assumption was not satisfied in the linear regression models with traumatic stress symptoms as an outcome, we recalculated the coefficient table using robust standard errors in those cases; this allowed for nonconstant variance. All correlational and regression analyses were conducted using R version 4.2.2.

Results

Sample characteristics

Three sets of data were used for analyses: (i) all participants (N = 139); (ii) a subset of participants who had experienced strangulation, irrespective of an AIC (n = 89); and (iii) a subset of participants who had experienced an AIC during strangulation (n = 42). The mean age of the women in the total sample was 40 years, with the majority being White. Only one woman reported residing at a shelter at the time of the interview (Table 1). Sixty-four percent of the sample indicated a history of strangulation, 30% indicated a history of a strangulation-related AIC, and 15% indicated a history of strangulation-related LOC. The mean average time since the most recent strangulation was 8.7 years (range = 2.5 weeks to 36.5 years; median = 7.2 years; interquartile range = 3.1–15.7 years).

Table 1.

Demographic Characteristics of All Study Participants, Participants Who Experienced IPV-Related Strangulation, Participants Who Experienced IPV Strangulation-Related AIC, and Participants Who Experienced IPV Strangulation-Related LOC

	Total sample	Experienced strangulation	Experienced strangulation-related AIC	Experienced strangulation-related LOC
Sample size	139	89 (64%)	42 (30%)	21 (15%)
Age (years), mean (SD)	40.0 (12.9)	40.9 (12.2)	40.8 (10.5)	40.9 (10.7)
Education (years), mean (SD)	15.6 (2.5)	15.5 (2.5)	15.3 (2.4)	15.8 (2.3)
Race, n (%)
White	98 (71)	63 (71)	33 (79)	16 (76)
Black/African American	17 (12)	11 (12)	4 (10)	3 (14)
Asian	7 (5)	5 (6)	1 (2)	0 (0)
Hawaiian/Pacific Islander	1 (1)	1 (1)	1 (2)	0 (0)
Multiracial	5 (4)	4 (4)	1 (2)	1 (5)
Other	11 (8)	5 (6)	2 (5)	1 (5)
Ethnicity, n (%)
Not Hispanic	112 (81)	69 (78)	36 (86)	19 (90)
Hispanic	26 (19)	19 (21)	6 (14)	2 (10)
Other	1 (1)	1 (1)	0 (0)	0 (0)
Residence, n (%)
Home	113 (81)	74 (83)	37 (88)	17 (81)
Shelter	1 (1)	1 (1)	1 (2)	1 (5)
Other	14 (10)	9 (10)	3 (8)	2 (9)
Missing	11 (8)	5 (6)	1 (2)	1 (5)

AIC, alteration in consciousness; LOC, loss of consciousness; IPV, intimate partner violence; “Other” residence, college dormitory, nursing home, hospital inpatient, etc.

Comparison of neurobehavioral and traumatic stress symptoms for women who had versus had not experienced strangulation

Table 2 summarizes the selected neurobehavioral symptoms for all study participants, stratified by strangulation. As explained above, the symptoms of interest included dizziness, vision problems, hearing difficulty, poor concentration, feelings of anxiety, and feelings of depression using the NSI and traumatic stress symptoms using the PCL-5 total score. Compared with women who had never experienced strangulation, women who had experienced strangulation reported significantly greater traumatic stress (p = 0.03) and vision problems (p < 0.01). When adjusted for relevant covariates, only the effect for vision problems remained statistically significant (p < 0.01; Table 3).

Table 2.

Between-Group Differences in Neurobehavioral Symptoms Stratified by Strangulation (n = 139), Strangulation-Related AIC (n = 89), and Strangulation-Related LOC (n = 42)

	Strangulation				Strangulation-related AIC				Strangulation-related LOC
Variable	Yes	No	Total	p (ES)	Yes	No	Total	p (ES)	Yes	No	Total	p (ES)
Total N (%)	89 (64.0)	50 (36.0)	139		42 (47.2)	47 (52.8)	89		21 (50.0)	21 (50.0)	42
PCL-5; mean (SD)	34.1 (18.9)	29.6 (17.5)	34.1 (18.9)	0.03 (0.18)	41.3 (19.4)	32.6 (18.2)	36.7 (19.2)	0.05 (0.21)	48.2 (14.3)	34.4 (21.6)	41.3 (19.4)	0.02 (0.35)
NSI symptoms; mean (SD)
Dizziness	0.7 (1.0)	0.6 (0.9)	0.6 (1.0)	0.90 (0.01)	0.9 (1.0)	0.4 (0.9)	0.7 (1.0)	0.01 (0.27)	1.1 (1.1)	0.7 (0.9)	0.9 (1.0)	0.19 (0.20)
Vision problems	1.0 (1.2)	0.3 (0.6)	0.7 (1.1)	<0.01 (0.32)	1.3 (1.1)	0.6 (1.1)	1.0 (1.2)	<0.01 (0.35)	1.5 (1.2)	1.1 (1.1)	1.3 (1.1)	0.33 (0.15)
Hearing difficulty	0.6 (1.0)	0.3 (0.6)	0.5 (0.9)	0.06 (0.16)	0.7 (1.1)	0.5 (0.9)	0.6 (1.0)	0.58 (0.06)	0.9 (1.2)	0.4 (1.0)	0.7 (1.1)	0.10 (0.25)
Poor concentration	1.9 (1.2)	1.6 (1.1)	1.8 (1.2)	0.09 (0.14)	2.3 (1.1)	1.6 (1.2)	1.9 (1.2)	<0.01 (0.28)	2.3 (1.1)	2.3 (1.1)	2.3 (1.1)	0.98 (0.00)
Feeling anxious or tense	2.0 (1.3)	1.9 (1.1)	2.0 (1.2)	0.64 (0.04)	2.2 (1.4)	1.9 (1.3)	2.0 (1.3)	0.31 (0.11)	2.3 (1.5)	2.1 (1.3)	2.2 (1.4)	0.75 (0.05)
Feeling depressed or sad	1.6 (1.3)	1.5 (1.1)	1.6 (1.2)	0.82 (0.02)	1.9 (1.4)	1.4 (1.1)	1.6 (1.3)	0.13 (0.16)	2.2 (1.4)	1.5 (1.3)	1.9 (1.4)	0.24 (0.18)

All significant p-values (≤0.05) are in bold.

ES, effect size; M, mean; N, sample size; SD, standard deviation; %, percentage; AIC, alteration in consciousness; LOC, loss of consciousness; NSI, Neurobehavioral Symptom Inventory. The item for vision problems is vision problems, blurring, or trouble seeing.

Table 3.

Multivariable Regression Models with Strangulation as an Explanatory Variable, Adjusting for Relevant Confounders (N = 139)

	Model 1 vision problemsa	Model 2 traumatic stressb
Variable	OR (95% CI)	Estimate (95% CI)
Strangulation	3.77 (1.61, 9.65) p < 0.01	4.60 (−2.12, 11.31) p = 0.18
Alcohol intake	1.01 (0.94, 1.06) p = 0.85	—
IPV-TBI history	2.14 (0.99, 4.76) p = 0.06	4.21 (−2.60, 11.01) p = 0.22
Neurological disorder	2.92 (1.04, 8.23) p = 0.04	13.16 (2.47, 23.84) p = 0.02
Traumatic stress	1.01 (0.99, 1.03) p = 0.36	—

All significant p-values (≤0.05) are in bold.

^aUses an ordinal logistic regression model.

^bUses a linear regression model.

IPV-TBI, intimate violence related traumatic brain injury; OR, odds ratio; CI, confidence interval; Alcohol Intake, Alcohol Use Disorder Identification Test (AUDIT) score.

Upon additional post hoc review of our data, we found that all women who reported “severe” or “very severe” vision problems had experienced IPV-related strangulation. Notably, all but one of these women reported that their vision problems had started either when the relationship started, during the relationship, after a particular abusive incident, or a few years after the last abusive incident. (One woman did not provide a time for the start of her vision problems).

Comparison of neurobehavioral and traumatic stress symptoms for women who had experienced strangulation with versus without an AIC

Compared with women who had experienced strangulation but had never experienced strangulation-related AICs, women who had experienced strangulation-related AICs reported greater traumatic stress (p = 0.05), dizziness (p = 0.01), vision problems (p < 0.01), and poor concentration (p < 0.01; Table 2). When controlling for potential covariates, a strangulation-related AIC was associated with vision problems (p < 0.01; Table 4).

Table 4.

Multivariable Regression Models with Strangulation-Related AIC as an Explanatory Variable, Adjusting for Relevant Confounders (N = 89)

	Model 1 dizzinessa	Model 2 vision problemsa	Model 3 poor concentrationa	Model 4 traumatic stressb
Variable	OR (95% CI)	OR (95% CI)	OR (95% CI)	Estimate (95% CI)
Strangulation-related AIC	2.23 (0.90, 5.71) p = 0.09	3.60 (1.56, 8.60) p < 0.01	2.14 (0.98, 4.73) p = 0.06	7.15 (−1.06, 15.35) p = 0.09
Neurological disorder	3.60 (1.07, 12.14) p = 0.04	3.24 (1.09, 9.80) p = 0.03	1.03 (0.30, 3.41) p = 0.96	7.92 (−2.79, 18.62) p = 0.15
TBI history	—	2.48 (0.99, 6.54) p = 0.06	—	5.51 (−3.80, 14.82) p = 0.24
Alcohol use	1.08 (1.00, 1.16) p = 0.04	—	1.12 (1.04, 1.22) p < 0.01	—
Traumatic stress	1.03 (1.00, 1.05) p = 0.03	—	1.05 (1.02, 1.07) p < 0.01	—

All significant p-values (≤0.05) are in bold.

^aUses an ordinal logistic regression model.

^bUses a linear regression model.

AIC, alteration in consciousness; TBI, traumatic brain injury; alcohol use, Alcohol Use Disorder Identification Test (AUDIT) score.

Comparison of neurobehavioral and traumatic stress symptoms for women who had experienced strangulation-related LOC versus those who had experienced an AIC

Compared with women who experienced strangulation-related AICs without LOCs, women who had experienced strangulation-related LOCs reported statistically significantly higher scores for traumatic stress symptoms (p = 0.02; Table 2). The association with traumatic stress remained significant after adjusting for relevant covariates (p = 0.02; Table 5).

Table 5.

Multivariable Regression Models with Strangulation-Related LOC as an Explanatory Variable, Adjusting for Relevant Confounders (N = 42)

	Model 1 hearing difficultya	Model 2 traumatic stressb
Variable	OR (95% CI)	Estimate (95% CI)
Strangulation-related LOC	4.48 (0.95, 15.05) p = 0.07	13.01 (1.95, 24.08) p = 0.02
Alcohol use	1.14 (1.02, 1.31) p = 0.02	—
TBI history	—	8.29 (−7.34, 23.92) p = 0.29
Neurological disorder	—	8.33 (−6.98, 23.64) p = 0.28

All significant p-values (≤0.05) are in bold.

^aUses an ordinal logistic regression model.

^bUses a linear regression model.

LOC, loss of consciousness; OR, odds ratio; TBI, traumatic brain injury; alcohol use, Alcohol Use Disorder Identification Test (AUDIT) score.

Discussion

Approximately two-thirds (i.e., 64%) of the women in our study, whom had all experienced IPV, reported a history of IPV-related strangulation; 30% reported strangulation-related alterations in consciousness, and 15% reported strangulation-related LOC. This is similar to previously published studies. In their study of women who had experienced IPV, Wilbur and colleagues reported that 68% of women reported having been strangled, and 17% reported strangulation-related LOC.¹³ Furthermore, Valera and Berenbaum noted that 27% of women who experienced IPV reported a strangulation-related AIC.⁸ In a four-country study (Canada, United States, Spain, Colombia) of women who experienced IPV, Adhikari and colleagues reported that 28% reported a strangulation-related AIC.¹⁷ The women in our sample were, apart from one, not living in a shelter and, on average, experienced their last instance of strangulation 8.7 years ago. Due to this long average time between strangulation events and the study interviews, if these symptoms began immediately or shortly after the strangulation event, the outcome measures of interest can thus be thought to represent possibly chronic neurobehavioral and traumatic stress symptoms.

We also found that women who were strangled reported higher levels of vision problems than women who were never strangled, even after adjusting for potential confounders, including IPV-related TBIs. Upon further examination, we found that all women who reported “severe” or “very severe” vision problems had experienced strangulation. All but one of these women reported that their vision problems had started either when the relationship started, during the relationship, after a particular abusive incident, or a few years after the last abusive incident. In addition, women who were strangled with an AIC reported higher levels of vision problems than women who were strangled without an AIC, even after adjustment for confounders. Other studies have reported changes to vision (e.g., blurred or impaired vision) in association with strangulation events, although as an acute symptom.^19–21,28 Acutely blurred vision and vision changes might be linked to strangulation through neurovascular compromise arising from the impairment of oxygenated blood flow due to pressure to the jugular veins, carotid arteries, and/or larynx.^40,41 However, given that the associations observed in these women are years later rather than during or immediately after the strangulation event, those links seem unlikely here. Future research is needed to better define, both objectively and subjectively, the vision difficulties experienced by these women and to identify the ophthalmological, visual-vestibular, neurological, and psychological factors that might be associated with these possible vision difficulties.

With respect to LOC relative to other AICs, our data show that women who experienced LOC during strangulation reported greater traumatic stress symptomatology than those who experienced other AICs during strangulation, even after adjustment for confounding effects. The traumatic stress findings are consistent with those of several studies showing an association between strangulation and psychological distress, including traumatic stress.^15,17,26,42 However, most of these studies focused on “choking” more generally, without considering AICs or LOCs, as we did here. In the present study, the effect size for traumatic stress was the largest in the women who reported experiencing LOC in association with the strangulation event. Our data are consistent with the findings of Karr and colleagues, who reported a positive relationship between the number of LOC episodes and greater emotional symptom severity (e.g., nervousness, temper outbursts).²⁵ Notably, Karr’s sample included women recruited specifically from courts after receiving a protective order against a partner, whereas our sample included women largely living in the community long after the abuse was over, irrespective of whether they sought assistance or protection for IPV.

We did not find strong associations between past strangulation and current symptoms of anxiety and depression in contrast to the results of other studies.^13,15,26 Although other studies have used full scales to obtain broader assessments of anxiety and depression, we used single-item questions about feeling anxious or depressed within the NSI. This methodological difference could explain, at least in part, why we did not find stronger associations. There are also a few associations worth considering for future studies, even though the associations were no longer statistically significant after adjusting for potential confounders. Considering our relatively small sample sizes—particularly when comparing LOC with other AICs—we lacked statistical power for our results pertaining to the associations between a strangulation-related AIC, dizziness, and poor concentration symptoms. Especially in light of other literature linking acute dizziness to strangulation^{13,19–21,27} and poor concentration to hypoxic-ischemic brain injuries, which may occur during strangulation,^16,22 future work with larger samples would be highly beneficial.

Our findings highlight the importance of considering a range of potentially confounding variables when trying to understand the long-term associations with IPV-related strangulation. The presence of possible neurological conditions, as defined by our study, was frequently associated with our variables of interest and also influenced our associations of interest. As such, it is important for studies to assess and address the presence of neurological conditions when trying to understand strangulation-related outcomes. It is also important to try to disentangle possible long-term outcomes associated with TBIs, alcohol use, drug use, and traumatic stress from possible outcomes associated with strangulation events. Finally, although we did not assess for instances of consensual strangulation, there is growing literature surrounding its high prevalence and potentially negative outcomes.^43,44 Future work should be informed by the literature on both these topics in trying to fully understand the outcomes of either consensual or nonconsensual strangulation.

This study has several limitations. First, there is no validated measure to assess for strangulation-related AICs or LOCs. This study used the BISA, which has been used extensively for the assessment of IPV-related AICs sustained from both blunt force trauma and strangulation.^8–11 However, the development of validated measures is important for future work in this area. Second, this study relies on self-reported strangulation events, which are subject to recall bias. Unfortunately, because most IPV, including strangulation, either goes unreported or untreated, relying on medical or police records or other data would be of little use in addressing these questions.⁴⁵ Third, our ability to consider the timing of confounding variables was limited by our study questions and also by our sample size. In the future, systematic and careful consideration of the timing of such confounding factors should be conducted in larger samples to permit analyses that would help us understand the complex relationships among these variables. Finally, our sample largely comprised White women living in Massachusetts. Additional research in this area is needed with women who have different sociodemographic characteristics and social determinants of health relating to race, ethnicity, cultural beliefs, gender identity, and living situation.

Strangulation is common among women subjected to IPV. We found that IPV-related strangulation was associated with both self-reported vision problems and traumatic stress years after the event. More research is needed on the short-, medium-, and long-term consequences of strangulation. Strangulation is under-reported, and more work is needed to encourage disclosure, social interventions, and timely health care for those subjected to these violent events.

Transparency, Rigor, and Reproducibility Statement

The study received approval from the Mass General Brigham Human Research Committee. The sample size was 139, and the data were acquired between July 2020 and December 2023; information on recruitment, participant inclusion–exclusion, and study design can be found in the Methods section. The average study visit was 2.5–3.5 h. Data collection was performed by trained test administrators, and all data collection instruments and software used can be readily accessed online. No data results were shared with individual participants. After the end of the study, the data will be made available in the Federal Interagency Traumatic Brain Injury Research repository. No code for statistical analyses used in this study will be made available. There are currently no planned replication or validation studies for this study. All the data were analyzed using R version 4.2.2. Statistical assumptions were satisfied for tests and regressions, except for constant variance in linear regression models with traumatic stress symptoms as an outcome. In that circumstance, the coefficient table was recalculated using robust standard errors to allow for nonconstant variance. All statistical analyses were performed by the first author and reviewed by the co-authors.