Utility of Aripiprazole in the Treatment of Anorexia Nervosa in Children and Adolescents: A Retrospective,Descriptive,Matched Cohort Study

Introduction

Anorexia nervosa (AN) is a life-threatening mental health condition characterized by weight loss achieved through dieting, fasting, and/or excessive exercise (restrictive type) or through self-induced vomiting, misuse of laxatives, or diuretics (binge-eating/purging type) (American Psychiatric Association, 2022). These behaviors are accompanied by severe cognitive distortions and obsessive preoccupations regarding eating, weight gain, food, and/or body image. While the neurobiology of AN is not completely understood, an imbalance of serotonin and dopamine may explain alterations in satiety and food-related reward/motivation (Frank et al., 2023; Marzola et al., 2015). Knowing that AN has the highest mortality rate of any mental health condition, timely identification and treatment are critical (American Psychiatric Association, 2022; Couturier et al., 2020; Crone et al., 2023; Hornberger et al., 2021).

While Family-Based Treatment (FBT) and weight restoration are mainstays of treatment among children and adolescents, interest exists in evaluating the role of atypical antipsychotics for severe cognitive distortions and obsessive preoccupations that often complicate treatment, interfere with participation in FBT, and/or worsen clinical outcomes (Frank, 2016; Frank et al., 2023, 2017). Among atypical antipsychotics, aripiprazole, olanzapine, quetiapine, and risperidone have been studied in youth with AN. Olanzapine is well-studied, with seven double-blind, randomized, controlled trials evaluating its use in youth or adults with AN (Attia et al., 2019; Bissada et al., 2008; Brambilla et al., 2007b, 2007a; Kafantaris et al., 2011; Mondraty et al., 2005). Results are mixed for weight gain and eating disorder (ED) cognitions, with notable increases in sedation, fasting blood glucose, liver function tests, prolactin, and total cholesterol. In a double-blind, placebo-controlled study, risperidone did not demonstrate improvements in core ED symptoms or time to reach 90% of ideal body weight (Hagman et al., 2011). Quetiapine did not demonstrate a difference in ED symptoms or body mass index (BMI) in the only double-blind, placebo-controlled study (Powers et al., 2012).

While olanzapine is the most widely studied atypical antipsychotic, evidence is mixed regarding its effectiveness, and high rates of side effects may contribute to frequent medication discontinuation (Attia et al., 2011, 2019; Bissada et al., 2008; Brambilla et al., 2007a, 2007b; Kafantaris et al., 2011 Spettigue et al., 2018). Additionally, risperidone and quetiapine have not demonstrated consistent benefit (Hagman et al., 2011; Powers et al., 2012). Aripiprazole has gained interest given its unique mechanism as a dopamine (D₂) receptor partial agonist and favorable side effect profile, including lower risk for sedation, metabolic side effects, and elevations in serum prolactin, in comparison to other atypical antipsychotics (Frank, 2016).

Current data to support the use of aripiprazole includes case series and a retrospective review suggesting positive outcomes on obsessive thoughts related to food, body image, fear of eating, weight restoration, and an increase in BMI/BMI percentile (Frank et al., 2017; Tahıllıoğlu et al., 2020; Thorey et al., 2023; Trunko et al., 2011). This study aims to evaluate the role of aripiprazole in children and adolescents with a high number of FABs that interfere with meal completion and progression through treatment.

Methods

ED treatment and participants

The EDP at Children’s Hospital Colorado is a day treatment program employing a modified version of FBT (i.e., caregivers in charge of food selection), highly structured meal support, multi-family groups, and treatment coordination with interprofessional child and adolescent psychiatry and adolescent medicine teams. Some youth are referred from the community, while others are admitted following stabilization of acute medical concerns. Target body weight is determined by a registered dietitian and the treatment team, based on a variety of clinical factors including review of growth curve trends and weight loss prior to admission. Weight trends are monitored throughout treatment, with caregivers adjusting nutrition to meet target weight goals.

Based on previously evaluated quality improvement data, most youth with anorexia nervosa—restricting type (AN-R) in this program are treated with thiamine, zinc, a multivitamin, and supportive gastrointestinal medications (e.g., constipation, nausea, heartburn treatment) on initial medical admission. Once admitted to the EDP, every patient is evaluated and followed by a psychiatric provider (child psychiatrist or advanced practice provider). Psychotropic medications may be recommended during treatment as an adjunct to FBT and the milieu-based group therapy, which utilizes the Unified Protocol.

Patients were included in this study if they had a diagnosis of AN and were started on aripiprazole for ED cognitions. Patients were excluded if they were started on aripiprazole for other indications than AN (e.g., major depressive disorder) or took aripiprazole for less than 10 days, based on previously published literature suggesting a lack of clinical effect during this time frame (Frank et al., 2017). Those who were treated with aripiprazole were matched 1:2 to those not treated with aripiprazole or another antipsychotic during the ED treatment program (Fig. 1).

OPEN IN VIEWER

FIG. 1.

Study flow.

Results

Study outcomes

Aripiprazole was associated with a reduction in FABs with a mean change from T1 to T3 of 3.5 versus 0.9 in the control group (p = 0.026) (Fig. 2). The mean increase in weight from T1 to T3 was greater among those prescribed aripiprazole, with a change of 4.8 kg compared with 3.1 kg among controls (p = 0.00024) (Fig. 3). Patient-specific percentage of target weight increased by 8.9% among those prescribed aripiprazole compared with 6.1% in the control group (p = 0.0016) (Fig. 4). Additionally, 42.9% of patients in the aripiprazole group discharged at or above target weight versus 29.8% of patients in the control group. The mean time to target weight was 41.2 versus 33.5 days for those prescribed aripiprazole and controls, respectively (p = 0.18).

OPEN IN VIEWER

FIG. 2.

Aripiprazole improves food avoidant behaviors (FABs). FAB scores are scored at each meal/snack during EDP, with a range of 0–40 (0 = best; 42 = worst). Data are represented as a daily average of FAB scores. Mean change in FAB scores from T1 to T3 was 3.5 (aripiprazole) vs. 0.9 (control group) (p = 0.026).

OPEN IN VIEWER

FIG. 3.

Aripiprazole-associated weight gain. The mean increase in weight from T1 to T3 was greater among those prescribed aripiprazole, with a change of 4.8 kg compared with 3.1 kg among controls (p = 0.00024).

OPEN IN VIEWER

FIG. 4.

Aripiprazole increases the likelihood of achieving target weight. Patient-specific percentage of target weight increased by 8.9% among those prescribed aripiprazole compared with 6.1% in the control group (p = 0.0016).

Concomitant medications were commonly prescribed, including a multivitamin, zinc, and gastrointestinal support (e.g., PPI/H2RA, simethicone, antinausea) (Table 2). Antidepressants were more often prescribed in those treated with aripiprazole (81% vs. 61.9%; p = 0.03). Additionally, antianxiety medications (e.g., lorazepam or hydroxyzine) were more often prescribed in the aripiprazole group (61.9% vs. 39.3%; p = 0.02).

OPEN IN VIEWER

Table 2.

Secondary Outcomes  

	Cases (n = 42)	Controls (n = 84)	p-Value
	n (%)	n (%)
Concomitant medications
MVI	42 (100)	84 (100)	1
Zinc	38 (90.5)	69 (82.1)	0.22
Antidepressant	34 (81)	52 (61.9)	0.03
SSRI	33 (78.6)	51 (60.7)	0.05
SNRI	1 (2.4)	1 (1.2)	0.06
Thiamine	12 (28.6)	20 (23.8)	0.67
GI support	42 (100)	80 (95.2)	0.15
Antianxiety	26 (61.9)	33 (39.3)	0.02
AD initiated during program	10 (23.8)	26 (31)	0.40
Time to target weight mean ± SD (range) (days)	41.2 ± 17.3 (14–71)	33.5 ± 19.6 (3–68)	0.18
Patients at target weight at program discharge	18 (42.9)	25 (29.8)	N/A

AD, antidepressant; GI, gastrointestinal; MVI, multivitamin; SD, standard deviation; SNRI, serotonin norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor.

Aripiprazole prescribing trends are described in Table 3. The mean starting dose was 1.9 mg/day with a mean discharge dose of 2.8 mg/day. Sedation and dizziness were most commonly reported (19%), with the majority (52.4%) not reporting any side effects. Overall, there was a low rate of monitoring for (14.3%) and metabolic labs (e.g., lipid panels) (26.2%). The majority (97.6%) of patients who were prescribed aripiprazole continued on the medication at discharge (Table 3).

OPEN IN VIEWER

Table 3.

Aripiprazole Prescribing

	n = 42, n (%)
Time of initiation, mean ± SD (range) (weeks)	3 ± 3 (0–10)
Starting dose, mean (range) (mg)	1.9 (1–5)
Dose at discharge, mean (range) (mg)	2.8 (0–5)
Frequency
Daily	3 (7.1)
Twice daily	27 (64.3)
At bedtime	12 (28.6)
Formulation
Tablet	40 (95.2)
Oral solution	2 (4.8)
Side effects
GI upset	3 (7.1)
Sedation	8 (19)
Dizziness	8 (19)
Headache	3 (7.1)
Other	6 (14.3)
None	22 (52.4)
Monitoring
Lipid panel	11 (26.2)
EPS	6 (14.3)
ECG	29 (69.0)
Total duration of use (range) (days)	36.5 (12–75)
Discharged from program with aripiprazole	41 (97.6)

ECG, electrocardiogram; EPS, extrapyramidal symptoms; GI, gastrointestinal; MG, milligrams; SD, standard deviation.

Discussion

Given the high utilization of psychotropic medications in individuals with AN, there is a need for more effective and better-tolerated alternatives. Our study evaluated the impact of aripiprazole in children and adolescents, 11–18 years of age, with AN admitted to an Eating Disorder Partial Hospitalization Program. Aripiprazole was typically initiated at program midpoint among patients with persistent cognitive rigidity, thought distortions, and obsessive preoccupations pertaining to food, and/or high FAB scores, which contributed to challenges in tolerating FBT and/or weight restoration. Starting and discharge doses of aripiprazole were low (≤5 mg/day) and most often administered twice daily to minimize the risk for nausea/vomiting (Findling et al., 2008). This differs significantly from previous studies that describe the use of higher doses (2.5–15 mg/day) (Tahıllıoğlu et al., 2020; Trunko et al., 2011).

Aripiprazole has demonstrated improvements in ED cognitions and weight gain in three case series (Frank, 2016; Tahıllıoğlu et al., 2020; Trunko et al., 2011) and one retrospective case-control study (Frank et al., 2017). The first published case series (N = 4 adolescent females) described improvements in cognitive flexibility and ED preoccupations surrounding weight gain after initiation of aripiprazole (2–5 mg/day). Self-reported improvement in symptoms was described within 1 week of aripiprazole initiation, with total duration of treatment ranging from 3 to 12 months (Frank, 2016). Among a group of nine patients who were previously treated with an antipsychotic, aripiprazole was associated with an increase in medication compliance, improved medication tolerability, and improved ED outcomes. Reasons for prior antipsychotic discontinuation included increased appetite with olanzapine and galactorrhea, sedation, or increased appetite with risperidone (Tahıllıoğlu et al., 2020). Increase in BMI, reductions in comorbid anxiety and depression, and reductions in fear of eating, obsessive ED thoughts, and cognitive rigidity were reported by eight females (15–55 years old) treated with aripiprazole 5–15 mg/day in a third case series (Trunko et al., 2011). Finally, in a retrospective, case-control study, aripiprazole (n = 22) was associated with a greater increase in BMI (6% higher) and BMI percentile (20% higher) compared with controls (n = 84) (Frank et al., 2017). While these studies are limited by small sample sizes and lack of rigorous study methodology, they provide important descriptive information regarding aripiprazole dosing, tolerability, and role in individuals with AN (Thorey et al., 2023).

Neurobiological mechanisms help explain why aripiprazole may be a more effective treatment strategy, compared with in-class alternatives (e.g., olanzapine). Current evidence suggests heightened responsiveness in brain reward circuits and hypersensitive dopamine systems as possible links to AN symptomatology (Frank, 2016; Frank et al., 2023, 2017). As a D₂ partial agonist, aripiprazole is thought to decrease and desensitize DA receptors over time, resulting in attenuation of reward system responsiveness, habituation to refeeding and fear extinction, reduced conditioned fear response, and reduced cognitive rigidity and body image distortions. Additionally, D₂ partial agonists are thought to improve learning and anxiety among malnourished females who are of low weight and in a low estrogen state (Frank, 2016; Frank et al., 2023, 2019).

While the optimal aripiprazole dosing strategy has not been identified for AN-associated cognitive distortions, obsessive preoccupations, and related behaviors, this and other studies highlight positive results with low doses (Frank, 2016; Frank et al., 2017). At low doses (≤5 mg/day), aripiprazole may predominantly target presynaptic D₂ autoreceptors and 5-HT1A receptors, functionally modifying DA release and sensitivity (de Bartolomeis et al., 2015; Tuplin and Holahan, 2017). Whereas, at higher doses, aripiprazole demonstrates high binding affinity for postsynaptic D₂ receptors, conferring a higher risk for adverse effects. The impact of this in AN has not been established and warrants further evaluation.

Notably, a diagnosis of MDD and treatment with an antidepressant were more common in those prescribed aripiprazole. Previously published data suggested similar trends, with selective serotonin reuptake inhibitors (SSRIs) frequently co-prescribed for the treatment of MDD predicting higher clinical severity of illness (Couturier et al., 2019; Riquin et al., 2021). Similar to our findings, other studies of aripiprazole in AN found that most adolescents were started on an antidepressant prior to initiation of ED treatment and/or before initiation of aripiprazole (Frank, 2016; Frank et al., 2017; Tahıllıoğlu et al., 2020; Trunko et al., 2011).

While this result might indicate an overall higher burden of mental health symptoms among the aripiprazole group, antidepressants are often prescribed in youth with AN targeting common comorbidities including MDD, GAD, PTSD, and/or OCD (Couturier et al., 2020; Hornberger et al., 2021; Lock, 2019). Notably, current evidence suggests antidepressants do not improve cognitive distortions among individuals with AN, with mixed results on BMI/BMI percentile (Chiu et al., 2023; Frank et al., 2017). While aripiprazole was prescribed for ED cognitions, secondary improvement in MDD cannot be ruled out.

Mean FAB score reductions were greater among those prescribed aripiprazole compared with controls (p = 0.026), suggesting improvement in distress tolerance, cognitive rigidity, thought distortions, and obsessive preoccupations pertaining to food. The mean change of 3.5 points observed in the aripiprazole group indicated improvement in time required to complete a meal or supplement, reduced distress completing a meal, and reduced urges to hide food, cut food into small pieces, or to avoid fear foods, as examples. The use of the EOES and change in observed FABs differs from other studies in which patient report (Frank, 2016; Frank et al., 2017; Trunko et al., 2011) and Clinical Global Impression—Severity scores (Tahıllıoğlu et al., 2020) have been used in previously published literature, highlighting the need for a more standardized, objective approach.

Our study demonstrated a larger mean change in weight among those prescribed aripiprazole compared with controls (4.8 vs. 3.1 kg, p = 0.00024). Additionally, treatment with aripiprazole increased the likelihood of achieving the target weight by program discharge. This is a particularly important finding, as current literature has predominantly described changes in BMI alone. Frank et al. (2017) describe modest improvement in BMI (6% higher at discharge) among those prescribed aripiprazole with clinically relevant improvements in BMI percentile (20% higher at discharge aripiprazole vs. no aripiprazole) among adolescents with AN. While BMI is an important clinical metric, it is well-established that body weight percentiles are more descriptive for children/adolescents as they are standardized to biological sex and age (Hornberger et al., 2021).

In addition to weight gain, time to achieve target weight and likelihood of achieving target weight are important clinical outcomes when considering the use of aripiprazole among adolescents with AN. The course of AN among pediatric patients is variable, with ∼50% achieving full recovery, ∼30% experiencing partial recovery, and ∼20% remaining chronically ill (Harrington et al., 2015; Mysliwiec, 2020). Physiological, psychological, and cognitive impairments associated with AN are significant in early adolescence, given the interruption of critical periods of physical, psychological, and social development (Couturier et al., 2020; Hornberger et al., 2021). While time to achieve target weight did not statistically differ among our cases and controls, youth prescribed aripiprazole were more likely to discharge at target weight (Table 2), highlighting an important functional and developmental outcome.

While rates of patient-reported adverse effects were low, it is important to acknowledge that metabolic and extrapyramidal side effect (EPS) monitoring was not routinely performed (Table 3). These findings may be in part due to ED treatment guidelines not providing unique monitoring recommendations for youth with AN, perceived differences in risk among this patient population, in addition to institution-specific practices (Couturier et al., 2020; Crone et al., 2023; Hornberger et al., 2021). Knowing that youth prescribed atypical antipsychotics are at an increased risk for developing metabolic side effects compared with adults, consideration should be made for routine metabolic monitoring (i.e., hemoglobin A1c and lipid panel) at baseline, 3 months, and 6 months, then annually. Additionally, EPS monitoring should be performed at baseline and every 6 months (Pringsheim et al., 2011).

Our study has several limitations. First, this was a single-center, retrospective, descriptive study that relied on chart review documentation, limiting generalizability. While this study demonstrates improvements in ED behaviors, larger prospective studies are needed to support widespread use of aripiprazole in youth with AN. Second, matching criteria could not account for all confounders in this retrospective review. The exposure time to aripiprazole versus the control group was not evaluated over the exact same time, which could impact results. Comorbidities and concurrent psychotropic medications were challenging criteria to match cases and controls based on the retrospective nature of our study design. Third, given high rates of comorbid MDD and antidepressant prescribing among those prescribed aripiprazole, clinical improvement cannot solely be attributed to the impact of aripiprazole on FABs. Acknowledging this limitation, it is difficult to control for these confounders given high rates of MDD and antidepressant prescribing in the general AN population. Finally, while a validation study is currently in progress, the EOES is not yet a validated tool at the time of article publication. However, the EOES has been used in our EDP since 2006 as a tool to standardize meal observations, determine the level of meal supervision, and identify patient-specific coaching strategies during meals.

More information is needed regarding the safety and efficacy of aripiprazole in children and adolescents with AN. Prospective, randomized, double-blind, placebo-controlled studies are needed to inform the ideal duration of aripiprazole treatment, persistence of aripiprazole effects on cognitive rigidity/distortions regarding food/body image, and long-term tolerability. Ideally, a future study would have patients discontinue other psychotropic medications and evaluate patients over the exact same time period.

While our study used FABs as an objective measure, a standardized rating scale is needed to support clinical decision-making and harmonize data reporting when evaluating atypical antipsychotic response in this patient population. A randomized, double-blind controlled trial is needed to strengthen generalizability and limit confounding impacts on the utility of aripiprazole in AN.

Notwithstanding these limitations, these findings add value to the available literature as the first matched cohort study evaluating the use of aripiprazole in children and adolescents with AN. Utilization of a standardized rating scale to measure improvement in cognitive rigidity and thought distortions provided a standardized, objective assessment of symptom improvement compared with previous studies that primarily relied on subjective patient report (Frank et al., 2017, 2017; Trunko et al., 2011). Finally, to our knowledge, this is one of the largest published reports evaluating the role of aripiprazole in youth with AN and to describe prescribing trends among this unique patient population.

Conclusion

Aripiprazole was associated with a decrease in FABs and an increased likelihood of achieving target weight in an ED treatment program. Additionally, low-dose aripiprazole was well-tolerated. This study provides additional support for the use of aripiprazole, among youth who continue to struggle with meal completion and emotional distress despite engagement in FBT, milieu-based therapy, and use of concomitant medications (e.g., SSRIs). Future studies should evaluate the optimal timing of aripiprazole initiation, dosing strategy, and ideal monitoring protocol.

Clinical Significance

Novel treatment strategies are needed to address core symptoms to improve treatment outcomes in youth with AN. Aripiprazole should be considered for youth with persistent severe cognitive distortions and obsessive preoccupations that interfere with weight restoration and have not improved with engagement in traditional treatment modalities (e.g., FBT).