Posttraumatic stress disorder, traumatic events, and longitudinal eating disorder treatment outcomes: a systematic review

A key limitation was that many included studies used measures without available or reported psychometric data.

“A notable limitation of the included studies was measurement methods that did not have available or reported psychometric properties.”

The other studies relied on clinical interviews whose psychometric properties were unpublished or unspecified.

“The remaining studies relied on clinical interviews with unpublished or unspecified psychometric properties.”

Among 12 studies of trauma exposure and eating disorder outcomes, only four used trauma measures with published psychometric evaluations.

“Of the 12 studies that examined the association between exposure to traumatic events and eating disorder treatment outcomes, less than a third of the studies (k = 4) used measures of traumatic events with published psychometric examinations.”

Treatment dropout was an outcome in four studies.

“Four studies examined treatment dropout as an outcome.”

One study found no dropout difference between people exposed to childhood trauma before eating disorder onset and those not exposed.

“Finally, one study did not find any difference between individuals that were exposed childhood traumatic events before the eating disorder onset and those that had not been exposed (Calugi et al., 2018).”

• cites Calugi et al., 2018Not checked yet

  Not checked in the free run.

Seven of those studies compared trauma-exposed and unexposed people at the end of treatment.

“Of these studies, seven examined differences at the end of treatment between individuals exposed to traumatic events compared with their unexposed peers.”

Three studies assessed only traumatic events that occurred before the eating disorder began.

“Furthermore, three studies limited their assessment of traumatic events to only before the onset of the eating disorder (Calugi et al., 2018; Carter et al., 2006; Rodríguez et al., 2005).”

• cites Calugi et al., 2018Not checked yet

  Not checked in the free run.

• cites Carter et al., 2006Not checked yet

  Not checked in the free run.

• cites Rodríguez et al., 2005Not checked yet

  Not checked in the free run.

At least six of the seven studies found no difference between trauma-exposed and unexposed people.

“Therefore, at least six of the seven studies examined found no differences between individuals exposed to traumatic events and unexposed peers.”

One study found dropout varied by diagnosis and trauma exposure: AN-BP patients with childhood trauma before eating disorder onset were more likely to stop treatment early than patients without trauma or AN-R patients with trauma.

“One study found that dropout rates depended on both diagnosis and traumatic event exposure, such that AN-BP patients with traumatic event histories in childhood before eating disorder onset were more likely to terminate treatment prematurely compared to individuals without traumatic events or AN-R patients with traumatic event histories (Carter et al., 2006).”

• cites Carter et al., 2006Not checked yet

  Not checked in the free run.

Eating disorder symptoms were examined as treatment outcomes in 11 studies.

“Eleven studies examined eating disorder symptoms as outcomes from treatment.”

That study did not report whether childhood sexual abuse alone related to post-treatment eating pathology differences.

“This study further did not report whether childhood sexual abuse was associated with differences in eating pathology post-treatment alone.”

Three of the four dropout studies found at least some link between trauma exposure and dropout.

“Therefore, three of four studies that examined treatment dropout as an outcome found some effect of traumatic events on dropout likelihood.”

Two studies reported that trauma exposure predicted more dropout from treatment.

“Of these studies, two found that traumatic events (childhood traumatic events, Castellini et al., 2018; any traumatic event history before eating disorder onset, Rodríguez et al., 2005) were associated with greater treatment dropout.”

• cites Castellini et al., 2018Not checked yet

  Not checked in the free run.

• cites Rodríguez et al., 2005Not checked yet

  Not checked in the free run.

Ridley et al. (2009) reported that people with childhood sexual abuse exposure showed more eating disorder symptoms at follow-up than peers without such exposure.

“Ridley et al. (2009) found that individuals exposed to childhood sexual abuse demonstrated higher eating disorder symptomology at follow-up than their unabused peers”

• cites Ridley et al. (2009)Not checked yet

  Not checked in the free run.

At post-treatment, most studies found no difference between trauma-exposed and unexposed groups.

“Furthemore, most studies that examined effects at the end of treatment found no difference between traumatic event exposure and non-exposed groups.”

Most studies, except one, found that people with trauma histories were more likely to leave treatment early than those without.

“The majority of studies (with one exception) indicated that individuals with traumatic event histories were more likely to drop out of treatment prematurely compared with those without such a history.”

Trauma exposure may predict more symptom relapse after treatment.

“Findings also indicated that traumatic events may be associated with greater symptom relapse post-treatment.”

Hazzard et al. (2021) found no significant main PTSD effect at six months.

“but there was no significant main effect of PTSD at the six-month follow-up.”

• cites Hazzard et al., 2021Not checked yet

  Not checked in the free run.

Rodríguez et al. (2005) reported that pre-onset trauma was tied to less early symptom remission.

“This study found that any traumatic event before eating disorder onset was significantly associated with reduced early symptom remission and an increased relapse rate (Rodríguez et al., 2005).”

• cites Rodríguez et al., 2005Not checked yet

  Not checked in the free run.

Trottier (2020) linked PTSD to premature termination.

“however, Trottier (2020) found that PTSD was associated with premature termination.”

• cites Trottier, 2020Not checked yet

  Not checked in the free run.

Five studies looked at whether PTSD diagnosis related to eating disorder symptom reduction during treatment.

“Five studies examined the reduction in eating disorder symptoms during treatment as a function of PTSD diagnosis”

Findings on PTSD and premature termination are inconclusive.

“Regardless, the results assessing the effect of PTSD on premature treatment termination are equivocal.”

Among reviewed studies, childhood trauma exposure predicted higher odds of hospitalization or re-hospitalization.

“Four studies found differences such that patients with childhood traumatic events were more likely to be hospitalized or re-hospitalized (K. P. Anderson et al., 1997; Castellini et al., 2018)”

• cites K. P. Anderson et al., 1997Not checked yet

  Not checked in the free run.

• cites Castellini et al., 2018Not checked yet

  Not checked in the free run.

Two included studies found no post-treatment differences between people exposed to sexual assault and those not exposed.

“Of these studies, two studies (Calugi et al., 2018; Luadzers, 1998) did not find differences after treatment between individuals that were exposed to sexual assault and those that were not.”

• cites Calugi et al., 2018Not checked yet

  Not checked in the free run.

• cites Luadzers, 1998Not checked yet

  Not checked in the free run.

Scharff, Ortiz, Forrest, Smith, et al. (2021) linked PTSD to more symptom recurrence at six months.

“but greater symptom recurrence at six-month follow-up.”

• cites Scharff, Ortiz, Forrest, Smith, et al., 2021Not checked yet

  Not checked in the free run.

Hazzard et al. (2021) found PTSD moderated the child abuse-binge frequency link: at six months, those with moderate/severe abuse plus lifetime PTSD had more binge episodes than those with moderate/severe abuse alone.

“Further, PTSD moderated the association between moderate/severe child abuse and binge frequency such that individuals who were exposed to moderate/severe child abuse and a lifetime PTSD diagnosis presented with greater binge frequency at the six-month follow-up as compared to individuals with moderate/severe child abuse that did not have a lifetime diagnosis of PTSD.”

• cites Hazzard et al., 2021Not checked yet

  Not checked in the free run.

Ten studies restricted diagnoses mainly to AN, BN, and BED, with ARFID and OSFED mostly absent.

“Similarly, most of the studies in this review (k = 10) only included AN, BN, and/or BED, largely excluding Avoidant-Restrictive Food Intake Disorder (ARFID) and Other Specified Eating Disorder (OSFED) diagnoses.”

Few studies looked at PTSD as a factor in eating disorder treatment.

“In addition, the number of studies examining PTSD as a factor in eating disorder treatment was small, indicating that more research should be conducted in this domain.”

In those five studies, patients generally showed significant post-treatment decreases in eating pathology.

“for all these studies, patients overall experienced significant decreases in eating pathology after treatment.”

Evidence suggests people with trauma histories may benefit as much from treatment as those without, despite baseline symptom differences.

“However, the evidence also suggests that individuals with traumatic event histories can equally benefit from treatment compared with those without a history of traumatic events, despite baseline differences in symptomology (Backholm et al., 2013; Scharff, Ortiz, Forrest, & Smith, 2021).”

• cites Backholm et al., 2013Not checked yet

  Not checked in the free run.

• cites Scharff, Ortiz, Forrest, & Smith, 2021Not checked yet

  Not checked in the free run.

Among four comparable studies, only one found a PTSD effect on end-of-treatment outcomes.

“Only one of the four studies with comparable designs found an effect of PTSD on outcomes at the end of treatment.”

Two studies found no link between PTSD diagnosis and post-treatment eating disorder outcomes.

“Two studies found no association between PTSD diagnosis and eating disorder treatment outcomes at post-treatment (Hicks, 2016; K. S. Mitchell, Singh, et al., 2021)”

• cites Hicks, 2016Not checked yet

  Not checked in the free run.

• cites K. S. Mitchell, Singh, et al., 2021Not checked yet

  Not checked in the free run.

After treatment ends, trauma exposure may relate to stronger return of eating disorder symptoms.

“Finally, traumatic events may be associated with greater eating disorder symptom resurgence after treatment ends.”

• AssumptionRests on an assumption: Posttraumatic and eating disorder symptoms have bidirectional functional associations.

People with a trauma history improved in treatment about as much as those without such a history.

“but individuals with a traumatic event history benefited from treatment similarly to their unexposed peers.”

Scharff, Ortiz, Forrest, Smith, et al. (2021) linked PTSD to larger eating disorder symptom reductions during treatment.

“Scharff, Ortiz, Forrest, Smith, et al. (2021) found that PTSD was associated with more substantial eating disorder symptom reductions while in treatment”

• cites Scharff, Ortiz, Forrest, Smith, et al., 2021Not checked yet

  Not checked in the free run.

The review's main conclusion is that trauma exposure seems linked to higher eating disorder treatment dropout.

“The main conclusion of this review is that traumatic events appear to be associated with greater eating disorder treatment dropout.”

Childhood trauma exposure was linked to higher relapse likelihood.

“and more likely to relapse (Carter et al., 2012).”

• cites Carter et al., 2012Not checked yet

  Not checked in the free run.

K. S. Mitchell, Singh, et al. (2021) and Masson et al. (2007) found no link between PTSD and treatment dropout.

“K. S. Mitchell, Singh, et al. (2021), and Masson et al. (2007) found that PTSD was not associated with dropping out of treatment”

• cites K. S. Mitchell, Singh, et al., 2021Not checked yet

  Not checked in the free run.

• cites Masson et al., 2007Not checked yet

  Not checked in the free run.

Both follow-up studies showed some PTSD effect on eating disorder outcomes.

“In contrast, both the studies that examined a follow-up period indicated some effect of PTSD on eating disorder outcomes.”

Rodríguez et al. (2005) reported that pre-onset trauma was tied to a higher relapse rate.

“This study found that any traumatic event before eating disorder onset was significantly associated with reduced early symptom remission and an increased relapse rate (Rodríguez et al., 2005).”

• cites Rodríguez et al., 2005Not checked yet

  Not checked in the free run.

No included study tested complex PTSD against eating disorder treatment outcomes.

“In addition, the current review found no studies that examined the association between “complex PTSD” (C-PTSD) – a condition characterized by repeated, ongoing exposure to interpersonally traumatic events beginning in childhood and symptoms such as identity and personality disturbances (Cloitre et al., 2009; Herman, 1992) – and eating disorder treatment outcomes.”

• cites Cloitre et al., 2009Not checked yet

  Not checked in the free run.

• cites Herman, 1992Not checked yet

  Not checked in the free run.

Among six studies without interpretive methodological problems, four linked childhood trauma to worse follow-up outcomes.

“Therefore, four of six studies without methodological differences hampering interpretation found that childhood traumatic events were associated with poorer outcomes during the follow-up period (K. P. Anderson et al., 1997; Carter et al., 2012; Castellini et al., 2018; Fallon et al., 1994).”

• cites K. P. Anderson et al., 1997Not checked yet

  Not checked in the free run.

• cites Carter et al., 2012Not checked yet

  Not checked in the free run.

• cites Castellini et al., 2018Not checked yet

  Not checked in the free run.

One study found no link between PTSD diagnosis and eating disorder outcomes at six-month follow-up.

“or six-month follow-up (K. S. Mitchell, Singh, et al., 2021).”

• cites K. S. Mitchell, Singh, et al., 2021Not checked yet

  Not checked in the free run.

Three studies reported that PTSD predicted or moderated treatment outcomes.

“Three studies found that PTSD predicted or moderated treatment outcomes.”

With longer post-treatment follow-up, people with trauma histories more often show symptom increases.

“This finding is attenuated by evidence suggesting that when longer follow-up periods are included after the termination of treatment, individuals with traumatic event histories are more likely to experience increases in symptomology (K. P. Anderson et al., 1997; Carter et al., 2012; Castellini et al., 2018; Fallon et al., 1994).”

• cites K. P. Anderson et al., 1997Not checked yet

  Not checked in the free run.

• cites Carter et al., 2012Not checked yet

  Not checked in the free run.

• cites Castellini et al., 2018Not checked yet

  Not checked in the free run.

Hazzard et al. (2021) reported that PTSD predicted more objective binge episodes at treatment end.

“Hazzard et al. (2021) found that PTSD predicted greater objective binge episode frequency at the end of treatment”

• cites Hazzard et al., 2021Not checked yet

  Not checked in the free run.

Trauma exposure predicted more dropout from eating disorder treatment.

“Results indicated that traumatic event exposure was associated with greater ED treatment dropout”

Childhood trauma exposure was linked to lower likelihood of symptom remission.

“less likely to achieve symptom remission (Castellini et al., 2018; Fallon et al., 1994)”

• cites Castellini et al., 2018Not checked yet

  Not checked in the free run.

• cites Fallon et al., 1994Not checked yet

  Not checked in the free run.

Studies of PTSD diagnosis and eating disorder symptom outcomes produced mixed results.

“Similarly, the results from studies examining the effect of PTSD diagnosis on eating disorder symptom outcomes were also mixed.”

PTSD and symptoms related to traumatic events may reduce the effectiveness of eating disorder treatment.

“posttraumatic stress disorder (PTSD) and symptoms associated with traumatic events may interfere with the efficacy of eating disorder treatment (Brewerton, 2007).”

• cites Brewerton, 2007Not checked yet

  Not checked in the free run.

In people diagnosed with an eating disorder, estimates of exposure to a traumatic event under DSM-5 PTSD criterion A range from roughly 18% to 100%.

“Estimates of traumatic event exposure vary widely, such that approximately 18–100% of individuals diagnosed with an eating disorder have been exposed to a traumatic event according to the DSM-5 PTSD criterion A definition (Backholm et al., 2013; Kjaersdam Telléus et al., 2021; K. S. Mitchell et al., 2012; Swinbourne et al., 2012).”

• cites Backholm et al., 2013Not checked yet

  Not checked in the free run.

• cites Kjaersdam Telléus et al., 2021Not checked yet

  Not checked in the free run.

• cites K. S. Mitchell et al., 2012Not checked yet

  Not checked in the free run.

Psychiatric comorbidity in general is linked to worse eating disorder treatment outcomes.

“Psychiatric comorbidity in general is associated with poorer treatment outcomes from eating disorders (Vall & Wade, 2015)”

• cites Vall & Wade, 2015Not checked yet

  Not checked in the free run.

Childhood traumatic events raise the probability of receiving an eating disorder diagnosis.

“Previous systematic reviews have found that childhood traumatic events (Caslini et al., 2016; Kimber et al., 2017; Molendijk et al., 2017; Pignatelli et al., 2017) and sexual assault (Chen et al., 2010; Madowitz et al., 2015; Smolak & Murnen, 2002) are associated with a higher likelihood of eating disorder diagnosis.”

• cites Caslini et al., 2016Not checked yet

  Not checked in the free run.

• cites Kimber et al., 2017Not checked yet

  Not checked in the free run.

• cites Molendijk et al., 2017Not checked yet

  Not checked in the free run.

Sexual assault raises the probability of receiving an eating disorder diagnosis.

“Previous systematic reviews have found that childhood traumatic events (Caslini et al., 2016; Kimber et al., 2017; Molendijk et al., 2017; Pignatelli et al., 2017) and sexual assault (Chen et al., 2010; Madowitz et al., 2015; Smolak & Murnen, 2002) are associated with a higher likelihood of eating disorder diagnosis.”

• cites Chen et al., 2010Not checked yet

  Not checked in the free run.

• cites Madowitz et al., 2015Not checked yet

  Not checked in the free run.

• cites Smolak & Murnen, 2002Not checked yet

  Not checked in the free run.

People whose trauma exposure occurred across the lifespan appeared to benefit from treatment as much as others.

“suggesting that individuals with traumatic events across the lifespan benefited equally from treatment.”

Current theory implies that eating disorder treatment could work less well when posttraumatic symptoms are not addressed concurrently.

“Therefore, the existing theory would suggest that eating disorder treatment may be less efficacious if posttraumatic symptoms are not addressed simultaneously.”

It is also unclear from mixed evidence whether comorbid PTSD or trauma exposure predicts more dropout from eating disorder treatment.

“Furthermore, evidence that comorbid PTSD or exposure to traumatic events is associated with greater dropout from eating disorder treatment is also mixed (e.g., Castellini et al., 2018; Trottier, 2020).”

• cites Castellini et al., 2018Not checked yet

  Not checked in the free run.

• cites Trottier, 2020Not checked yet

  Not checked in the free run.

People with eating disorders who have been exposed to traumatic events or who have PTSD show more severe psychopathology, including greater eating, depressive, and anxiety symptoms, than those without such exposure.

“Exposure to traumatic events and/or a diagnosis of PTSD is also associated with more severe psychopathology in individuals with eating disorders, such as greater eating, depressive, and anxiety symptoms as compared to individuals who have not been exposed to traumatic events (Backholm et al., 2013; Scharff, Ortiz, Forrest, & Smith, 2021).”

• cites Backholm et al., 2013Not checked yet

  Not checked in the free run.

• cites Scharff, Ortiz, Forrest, & Smith, 2021Not checked yet

  Not checked in the free run.

A possible functional connection may exist between distressing emotions related to PTSD and eating disorder symptoms used as self-medication.

“indicating a potential functional link between uncomfortable emotional experiences associated with PTSD symptoms and eating disorder symptoms as self-medication (Brewerton, 2011).”

• cites Brewerton, 2011Not checked yet

  Not checked in the free run.

Unless posttraumatic beliefs are addressed, three mechanisms that maintain eating disorders would continue.

“Therefore, without dismantling these posttraumatic beliefs, three important maintaining mechanisms for eating disorders would persist.”

People may use eating disorder symptoms as a way to manage negative emotions.

“indicating that eating disorder symptoms may be used to cope with negative affect.”

Several cross-sectional analyses have shown that problems with emotion regulation mediate the relationship between trauma exposure or PTSD and eating pathology.

“Indeed, multiple cross-sectional analyses have found that emotion regulation difficulties mediate the association between traumatic events and/or PTSD and eating pathology (Mills et al., 2015; K. S. Mitchell & Wolf, 2016; Moulton et al., 2015; Racine & Wildes, 2015)”

• cites Mills et al., 2015Not checked yet

  Not checked in the free run.

• cites K. S. Mitchell & Wolf, 2016Not checked yet

  Not checked in the free run.

• cites Moulton et al., 2015Not checked yet

  Not checked in the free run.

Qualitative studies found that people with PTSD and eating pathology said they ate to manage irritability tied to their assault.

“For example, individuals with PTSD and eating pathology in qualitative studies have endorsed eating to manage the irritability associated with their assault (Breland et al., 2018).”

• cites Breland et al., 2018Not checked yet

  Not checked in the free run.

Differences in measurement may have affected the effects seen across studies.

“Measurement differences could therefore have influenced the observed effects across studies.”

The included papers tended to be of weak quality.

“Additionally, the quality of the papers in this review was skewed toward weak quality.”

People with PTSD may gain from treatment similarly to people without PTSD.

“individuals with PTSD may have similar treatment gains compared to individuals without PTSD”

Because diagnoses were mixed and not split by symptom cluster, possible effects may have been hidden.

“Further, the heterogeneity in examined diagnoses, without separating by symptom cluster, may have obscured potential effects.”

Rodríguez et al. (2005) could not be easily compared with studies covering full treatment or post-treatment follow-up.

“Rodríguez et al. (2005) was therefore difficult to compare to studies that examined a full course of treatment or follow-up periods after treatment.”

• cites Rodríguez et al., 2005Not checked yet

  Not checked in the free run.

The review may therefore overrepresent more severe and recent child abuse.

“Thus, this review may have been biased toward assessing more severe and relatively recent child abuse.”

People with PTSD may relapse more after treatment.

“but individuals with PTSD may experience greater symptom relapse post-treatment.”

Ridley (2009) and Vrabel et al. (2010) are hard to interpret because their reporting and analysis methods differ.

“For two studies (Ridley, 2009; Vrabel et al., 2010), interpretation is difficult due to methodological differences in reporting and analysis.”

• cites Ridley, 2009Not checked yet

  Not checked in the free run.

• cites Vrabel et al., 2010Not checked yet

  Not checked in the free run.

Ridley et al. (2009) failed to adjust for eating pathology scores before treatment or at termination.

“however, this study did not control for pre-treatment or termination eating pathology scores.”

• cites Ridley et al. (2009)Not checked yet

  Not checked in the free run.

Requiring binary measures means the findings cover only binary trauma/PTSD indicators and eating disorder outcomes.

“However, by imposing dichotomous measurement requirements, the current results are limited to only the association between binary indicators of traumatic events and PTSD with eating disorder outcomes.”