Treatment-Resistant Depression in Adolescents: Clinical Features and Measurement of Treatment Resistance

Abstract

Objective: To describe the clinical characteristics of adolescents with antidepressant treatment-resistant major depressive disorder (MDD) and to examine the utility of the Antidepressant Treatment Record (ATR) in categorizing treatment resistance in this population.

Methods: Adolescents with treatment-resistant MDD enrolled in an interventional study underwent a baseline evaluation with the ATR, Children's Depression Rating Scale-Revised (CDRS-R), and Clinical Global Impressions-Severity (CGI-S) scales. Demographic and clinical characteristics were examined with regard to ATR-defined level of resistance (level 1 to ≥3) using analysis of variance and χ² tests.

Results: In adolescents with treatment-resistant MDD (N = 97), aged 12–21 years, most were female (65%), white (89%), and had recurrent illness (78%). Patients were severely ill (median CGI-S score of 5), had a mean CDRS-R score of 63 ± 10, and 17.5% had been hospitalized for depression-related symptoms. Fifty-two patients were classified as ATR 1, whereas 32 were classified as ATR level 2 and 13 patients as ≥3, respectively. For increasing ATR-defined levels, illness duration increased from 12.0 (range: 1.5–31.9) to 14.8 (range: 1.8–31.7) to 19.5 (range: 2.5–36.2) months and the likelihood of treatment with serotonin norepinephrine reuptake inhibitors (SNRIs) and dopamine norepinephrine reuptake inhibitors (DNRIs) similarly increased (p = 0.006 for both SNRIs and DNRIs) as did the likelihood of treatment with mixed dopamine serotonin receptor antagonists (χ² = 17, p < 0.001).

Conclusions: This study underscores the morbidity and chronicity of treatment-resistant MDD in adolescents. The present characterization of related clinical features describes the use of nonselective serotonin reuptake inhibitors in adolescents with treatment-resistant depression and raises the possibility that those with the greatest medication treatment resistance are less likely to have had recurrent episodes. The study also demonstrates the utility of the ATR in categorizing treatment resistance in adolescents with MDD.

Introduction

As many as 40% of adolescents with major depressive disorder (MDD) fail to respond to initial pharmacotherapy (e.g., selective serotonin reuptake inhibitor [SSRI]) or psychotherapy) (Emslie et al. 2002, 2009). Furthermore, less than a third of adolescents reach remission with initial treatment (March et al. 2004, 2007; Curry et al. 2011). Youth with treatment-resistant depression (TRD) are more likely to attempt suicide, have impaired peer and family relationships, and incur additional psychiatric comorbidity. Despite the prevalence of treatment failure in adolescents with MDD, there are no validated approaches for characterizing the degree of treatment resistance in adolescents (Strawn and Croarkin 2020). Furthermore, the demographic and clinical characteristics of adolescents with TRD are largely unknown.

The largest examination of adolescents with TRD, the Treatment of SSRI-Resistant Depression in Adolescents (TORDIA) Study, randomized adolescents aged 12–17 (N = 334) (March et al. 2004). In this study of adolescents aged 12–18 years, patients were required to have failed treatment with an SSRI and to have been treated for at least 8 weeks. Moreover, SSRI treatment had to have been >40mg/d of fluoxetine or its equivalent (e.g., 40 mg paroxetine, 40 mg citalopram, 20 mg escitalopram, 150 mg sertraline). Youth who could not tolerate an SSRI dose equivalent to 40 mg/d of fluoxetine were also eligible. In the TORDIA study, most patients were white and female (>60%). Patients also had significant comorbidity, including anxiety disorders, attention-deficit/hyperactivity disorder, and posttraumatic stress disorder. Moreover, most patients had adolescent-onset of their depressive symptoms, experienced ∼2 years of depressive symptoms, and more than one in five had a prior history of a suicide attempt (Brent et al. 2008). However, the degree of treatment resistance and the way in which the patients differed across levels of treatment resistance was not formally characterized in the TORDIA study.

In adolescents with MDD, no studies have systematically categorized the degree of antidepressant treatment resistance. By contrast, in adults, antidepressant treatment resistance has been examined and characterized with measures that quantify treatment resistance. In the first attempt to categorize treatment resistance in adults with TRD, Thase and Rush (1997) proposed a five-staged assessment strategy with higher levels reflecting greater treatment resistance. More recently, the Committee for Medicinal Products for Human Use (CHMP) defined TRD based on “consecutive treatment with two products of different pharmacological classes, used for a sufficient length of time at an adequate dose” that fail to result in response or “a clinically meaningful effect” (CHMP 2013). These approaches require a number of assumptions and operationalized definitions. What constitutes a different pharmacological class (e.g., dopamine norepinephrine reuptake inhibitor [DNRI] vs. serotonin norepinephrine reuptake inhibitor [SNRI])? What suffices as an adequate duration of treatment? What is a minimum dose for each medication in this population? Should a minimum duration of medication exposure be stipulated and measured? How is psychotherapy classified with regard in terms of its equivalence to medication treatment? The various methods of classifying medication treatment resistance have been recently reviewed (Trevino et al. 2014). There have been divergent approaches, but, in recent years, the Antidepressant Treatment History Form (ATHF) has become the most common instrument to categorize the degree of treatment resistance in adults and has been used in studies of electroconvulsive therapy, transcranial magnetic stimulation (TMS), TRD, and sequenced treatment trials of antidepressants (Sackeim 2001; Rush et al. 2006). This tool evaluates prior antidepressant treatments in terms of dose, duration, and adequacy of adherence and clinical outcome. In its original form, this tool allows a measure of categorical number of treatments that failed to provide benefit, as well as measuring the “potency” of the treatment trial (i.e., in terms of dose used and duration of use). This tool has become important as it is one of the few antidepressant resistance inventories that has shown prospective validity in predicting future treatment response (Prudic et al. 1990; Sackeim et al. 1990). As a component of the original development of the NeuroStar device, the ATHF methodology was adapted into a simplified for use in nonresearch clinical settings in the form of the Antidepressant Treatment Record (ATR) (Mark A. Demitrack, MD, personal communication 2020). The ATR is sensitive and specific in its ability to identify ATHF treatment resistance levels ranging from 1 through 4 in adults with MDD (Neuronetics, data on file). However, the two instruments vary considerably. The ATHF requires substantially more time to complete than the ATR and prescribes a minimum effective doses and treatment durations for medications. For tricyclic antidepressants and lithium (adjunctive), the ATHF also documents blood levels to establish a minimum exposure. For tolerability, the ATHF classifies patients as intolerant if they cannot receive an “adequate” trial (based on dose or duration) or “if the dose required for response has increased due to reduction of response over time and the patient can no longer tolerate an effective dose,” whereas the ATR does not. The ATHF recommends a record review versus patient-reported history to establish higher “confidence” scores, whereas the ATR does not. For these reasons, the ATR potentially represents a more easily implemented assessment in the clinic.

To date, neither the ATHF nor ATR have been systematically studied with regard to their ability to characterize antidepressant treatment resistance in adolescents with TRD. With these considerations in mind this study sought to evaluate demographic and clinical characteristics of youth at various ATR levels of treatment-resistant MDD who were enrolled in an interventional study for TRD.

Methods

This study (ClinicalTrials.gov Identifier: NCT02586688) was approved by a central Institutional Review Board (IRB) or a site-specific IRB at each of the 14 sites. Of the sites, eight were academic (Mayo Clinic, University of Cincinnati, UCLA, Stanford University, Medical University of South Carolina, Bradley Children's Hospital/Brown University, Sheppard Pratt and The Ohio State University) and six were community-based clinical research organizations (Dothan Behavioral Medicine, Dothan, Alabama; Rocky Mountain TMS, Orem, Utah, Florida Clinical Practice Association, Inc., Gainesville, Florida, Anchor Neuroscience, Pensacola, Florida; Beacon Medical Group, South Bend Indiana and Integrative Psychiatry, Louisville, Kentucky). Written informed consent and assent were obtained from parents/guardians and from patients before inclusion in the study.

Subjects

Study participants were outpatients aged 12 through 21 years who met DSM-5 criteria for MDD, as assessed by the Mini-International Neuropsychiatric Interview for Children and Adolescents (MINI Kid) for participants ages 12–17 years or the MINI Neuropsychiatric Interview for participants ages 18–21 years. Participants were required to have a current episode of MDD lasting at least 4 weeks but ≤3 years in duration. Participants had resistance to antidepressant medication treatment in a discrete episode of MDD, current or prior episode if no treatment was given in the current episode.

Measures

Baseline data collected included demographics and clinical characteristics (e.g., comorbidity, duration of episode). Depressive symptom severity was assessed with the Children's Depression Rating Scale-Revised (CDRS-R), a 17-item interview that ranges in total score from 17 to 113, with a score of ≥40 consistent with clinically significant depression (Mayes et al. 2010). The Clinical Global Impressions-Severity (CGI-S), a clinician-administered instrument, was administered by study clinicians and is rated on a 7-point scale (Guy 1976). Scores of 1 reflect patients who are “normal, not at all ill”; scores of 2 reflect patients who are “borderline mentally ill”; scores of 3 describe patients who are “mildly ill”; scores of 4 reflect patients who are “moderately ill”; scores of 5 reflect patients who are “markedly ill”; and scores of 6 and 7 are associated with the descriptions “severely ill” and “among the most extremely ill patients,” respectively. The rating—which is performed by a clinician—is based both on observed and reported symptoms, functional impairment, and behavior for a 7-day period. The ATR was completed by a psychiatrist after evaluation of the patient.

Statistical analysis

CDRS-R scores and CGI scores were evaluated descriptively with regard to ATR treatment resistance stratification (e.g., ATR 1, ATR 2, ATR ≥3) and were statistically evaluated with an analysis of variance (ANOVA), as previously utilized in the comparison of symptom rating scales in pediatric patients with affective and anxiety disorders (Patel et al. 2006; Mossman et al. 2017). In addition, demographic and clinical variables were compared between ATR groups using ANOVA or χ² tests, as appropriate. All analyses were conducted in Statistical Analysis System (SAS, SAS Institute for Advanced Analytics, version 9.4). p-Values <0.05 were considered statistically significant and, given the exploratory nature of these analyses, no correction for multiple comparisons was made.

Results

Demographic and clinical characteristics of adolescents with treatment-resistant MDD

Of 112 patients randomized in this trial, 97 met criteria for treatment-resistant MDD definition. Adolescents with TRD (N = 97) were ∼17 years of age and most were female (65%), white (89%), and had recurrent illness (78%). Patients were severely ill as reflected by a median CGI-S score of 5 and a mean CDRS-R score was 63 ± 10 and 17.5% of patients had a prior depression-related hospitalization.

ATR classification of adolescents with MDD

Fifty-two patients were classified as ATR 1, whereas 32 and 13 patients were classified as ATR level 2 and ≥3, respectively. For these three ATR-defined levels, illness duration increased from 12.0 (range: 1.5–31.9) to 14.8 (range: 1.8–31.7) to 19.5 (range: 2.5–36.2) months for the three groups. Age distribution was similar across the three ATR classifications (χ² = 5.9, p = 0.21) (Fig. 1) and these ATR groups did not differ in terms of race (χ² = 6, p = 0.43, Table 1), ethnicity (χ² = 0.9, p = 0.63), or gender (Table 1).

FIG. 1.

Patient age ranges within each antidepressant treatment record (ATR) classification.

Table 1.

Demographic and Clinical Characteristics of Adolescents with Treatment-Resistant Major Depressive Disorder

	ATR 1, n = 52	ATR 2, n = 32	ATR ≥ 3, n = 13	Statistic
Age (years), mean ± SD	17.1 ± 2.6	17.7 ± 1.8	17.6 ± 1.8	χ2 = 5.9, p = 0.21
12–14	13 (25.0)	2 (6.3)	2 (15.4)
15–17	18 (34.6)	15 (46.9)	7 (53.8)
18–21	21 (40.4)	15 (46.9)	4 (30.8)
Race				χ2 = 6.0, p = 0.43
White	44 (84.6)	29 (90.6)	13 (100)
Black/African American	3 (5.8)	3 (9.4)	0
Asian	4 (7.7)	0	0
Other	1 (1.9)	0	0
Ethnicity (% Hispanic or Latino)	2 (3.8)	2 (6.3)	0 (0)	χ2 = 0.9, p = 0.63
Gender (% female)	35 (67.3)	20 (62.5)	8 (61.5)	χ2 = 0.3, p = 0.87
Primary diagnosis of MDD, n (%)	52	32	13	χ2 = 6.9, p = 0.032
Single episode	7 (13.5)	8 (25)	6 (46.2)
Recurrent episodes	45 (86.5)	24 (75)	7 (53.8)
Duration of current episode (months)				χ2 = 2.9, p = 0.258
Median	12.0	14.8	19.5
Minimum–maximum	1.5–31.9	1.8–31.7	2.5–36.2
<24 Months, n (%)	48 (92.3)	29 (90.6)	10 (76.9)
≥24 Months, n (%)	4 (7.7)	3 (9.4)	3 (23.1)
Secondary psychiatric diagnosis, N (%)	30 (57.7)	21 (65.6)	6 (46.2)	χ2 = 1.5, p = 0.473
ADHD	7 (13.5)	4 (12.5)	1 (7.7)	χ2 = 0.3, p = 0.852
Anxiety disorder	25 (48.1)	18 (56.3)	6 (46.2)	χ2 = 0.6, p = 0.725
Prior treatment history, n (%)a
Selective serotonin reuptake inhibitor	44 (84.6)	30 (93.8)	11 (84.6)	χ2 = 1.7, p = 0.438
Serotonin norepinephrine reuptake inhibitor	10 (19.2)	13 (40.6)	8 (61.5)	χ2 = 10, p = 0.006
Dopamine norepinephrine reuptake inhibitor	12 (23.1)	12 (37.5)	9 (69.2)	χ2 = 10, p = 0.006
Mixed dopamine serotonin receptor antagonist	3 (5.8)	12 (37.5)	6 (46.2)	χ2 = 17, p < 0.001
Electroconvulsive therapy	0 (0)	0 (0)	0 (0)
Hospitalization for depression	12 (23.1)	4 (12.5)	1 (7.7)	χ2 = 2.5, p = 0.281
Psychotherapy	10 (19.2)	6 (18.8)	1 (7.7)	χ2 = 1.0, p = 0.604
Symptom severity
CDRS-R,b mean ± SD	62.2 ± 9.8	62.7 ± 9.9	63.8 ± 11.4	F2, 94 = 0.1, p = 0.89
Median	63	64	64
Minimum–maximum	41–89	40–82	45–84
CGI-S, mean ± SD	5.2 ± 0.7	5.1 ± 0.7	5.0 ± 0.7	F2, 96 = 0.6, p = 0.58
Median	5	5	5
Minimum–maximum	4–7	4–6	4–6

^aCurrent or past episode.

^bOne subject was missing a baseline CDRS-R score in the ATR-2 and ATR3 groups.

ADHD, attention-deficit/hyperactivity disorder; ATR, Antidepressant Treatment Record; CDRS-R, Children's Depression Rating Scale-Revised; CGI-S, Clinical Global Impressions-Severity; MDD, major depressive disorder; SD, standard deviation.

Clinical severity, reflected by CDRS-R score and CGI-S, was similar among ATR groups (Table 1), whereas the number of patients with single (vs. recurrent episodes) significantly differed by ATR classification (χ² = 6.9, p = 0.032); more patients classified as ATR 1 and 2 had recurrent episodes compared with those classified as ATR ≥3. This was balanced by significantly more patients with ATR ≥3 having episodes ≥24 months (23% compared with 9.4% of patients with ATR 2 and 7.7% of patients with ATR 1, Table 1).

ATR classification of prior treatment

In terms of prior pharmacotherapy, most patients across ATR groups had received an SSRI (range 85%–94%) and prior treatment with SSRIs did not differ among the ATR groups (χ² = 1.7, p = 0.438). However, for “next line” interventions, SNRIs (χ² = 10, p = 0.006, Table 1), DNRI (χ² = 10, p = 0.006), and mixed dopamine serotonin receptor antagonists χ² = 17, p < 0.001, Table 1) were used more commonly in patients with greater treatment resistance. Prior treatment with psychotherapy did not differ among ATR categories (χ² = 1.0, p = 0.604, Table 1) and ranged from 8% to 19%.

Discussion

This study underscores the significant morbidity and chronicity of TRD in adolescents while characterizing the demography and clinical features of TRD in adolescents. In fact, it is one of only two studies (Brent et al. 2008) to examine the demographic and clinical characteristics of youth with TRD. Furthermore, this study is the first to demonstrate that the ATR can be successfully utilized in a large sample of adolescents with TRD. Several of our findings warrant additional discussion as they raise important issues with regard to treatment, classification, and heterogeneity of TRD in adolescents.

Our patients were remarkably similar to the only other large examination of the phenomenology of TRD in adolescents with two notable exceptions. First, few of our patients had prior psychotherapy treatment (18%) compared with TORDIA in which most patients had prior psychotherapy. Second, comorbidity may have been lower in this sample compared with TORDIA, although this may be related to the assessment of comorbidity and exclusion criteria of the two studies. The semistructured Kiddie Schedule for Affective Disorders and Schizophrenia was used in the TORDIA study, whereas this study used the structured MINI Kid (Sheehan et al. 2010) or MINI Neuropsychiatric Interview (Lecrubier et al. 1997). The structured assessment, with less direct questioning of parents, may have attenuated the ability to detect comorbidity. Importantly, capturing diagnostic and clinical heterogeneity, including comorbidity and prior multimodal treatment exposure, is critical in personalizing treatment. Specific clinical characteristics (e.g., number of comorbid disorders) may predict or moderate treatment response in adolescents with SSRI-resistant MDD as well in adolescents with MDD, anxiety disorders, and may even influence placebo response. Specifically, in TORDIA, the superiority of cognitive behavioral therapy (CBT) and CBT+medication was more pronounced in youth who had greater comorbidity (Asarnow et al. 2009). Finally, in our sample and in TORDIA (Brent et al. 2008), >80% of patients were Caucasian. This raises the possibility that non-Caucasians with TRD might be less likely to seek treatment or may be less likely to access treatment. If true, this is concerning given that race may affect treatment response and affects SSRI tolerability in youth with depressive and anxiety disorders (Ramsey et al. 2019).

In our sample, prior psychopharmacologic treatment was diverse. Many patients received evidence-based interventions, yet others received interventions that were scantly, if at all, supported by evidence in adolescents with MDD. The ATHR (the predecessor of the ATR) and ATR are based on medications and augmentation strategies for adults with MDD and do not consider differential medication tolerability in pediatric patients (e.g., venlafaxine being associated with greater treatment-emergent suicidality compared with SSRIs) (Brent et al. 2009; Mills and Strawn 2020) or differential efficacy (e.g., paroxetine in adults vs. youth with MDD) (Liu et al. 2019; Strawn et al. 2019a). In addition, the ATHR and ATR evaluate mediation dose; however, in pediatric populations some SSRIs have nonlinear kinetics and pharmacogenetic differences produce in significant variability in medication exposure. In this regard, recent studies suggest that for adolescents who are poor CYP2C19 metabolizers, 50 mg of sertraline and 10 mg of escitalopram are equivalent to 225 and 30 mg, respectively, in ultrarapid CYP2C19 metabolizers (Strawn et al. 2019b).

Limitations

Although this is the first study to prospectively examine the ATR in adolescents with TRD, several limitations warrant discussion. First, the MINI Kid and MINI may have decreased our ability to identify comorbidity. Second, while we focused on quantifying medication resistance, a focused measure or approach may have captured more information on prior psychotherapeutic treatment. Third, we did not capture objective measures of combined treatment (e.g., SSRI + stimulant + mixed dopamine serotonin receptor antagonist), which are common and poorly understood clinical approaches in adolescents with TRD. Fourth, these analyses are exploratory; the study eligibility required participants to have failed at least one medication in the current episode. The study was not originally designed or powered to detect or evaluate differences across ATR groups. Fifth, we lack data for the Antidepressant Treatment Response Questionnaire (ATRQ), which is increasingly used in clinical trials (Chandler et al. 2010). The ATRQ is a self-report measure that may have utility in adolescents and assesses the current episode.

Conclusion

The present findings are timely given that clinicians increasingly utilize novel off-label intervention for adolescent TRD such as ketamine (Cullen et al. 2018) and TMS (Croarkin et al. 2018). A systematic and valid approach for quantifying treatment resistance is critical. This approach will assist in identifying treatment-responsive phenotypes for interventional psychiatry procedures such as ketamine and brain stimulation to guide shared decision-making with families. Furthermore, prospective studies of these interventions often take years to complete and clinical practice outpaces the related evidence base. A harmonized approach to staging treatment resistance in adolescents could facilitate larger retrospective registry studies that will likely play a key role in understanding these newer modalities. Finally, as clinicians utilize newer interventions for adolescents with TRD, we must develop consensus definitions for the minimum duration and dose for “adequate” treatment trials.

Clinical Significance

This is the first study to examine the differential findings among ATR level of treatment resistance with respect to demographics and clinical features in adolescents with TRD. This study highlights the importance and utility quantifying treatment resistance in adolescents with TRD. The ATR appears to have utility and validity in this context.

Disclosures

Dr. Strawn has received research support from Allergan, Neuronetics, Otsuka, the National Institute of Mental Health, Eunice Kennedy Shriver National Institute of Child Health and Development, and National Institute of Environmental Health Sciences as well as the Yung Family Foundation. He receives royalties from Springer Publishing and received honoraria from CMEology and Neuroscience Educational Institute. He receives royalties from UpToDate. Finally, he received material support from and provided consultation to Myriad Genetics. Dr. Aaronson has received research support from Compass Pathways and Neuronetics and serves as a consultant to LivaNova, Neuronetics, Janssen, Genomind, and Sage Therapeutics. He serves on speaker boards for Janssen and Sunovion. Dr. Holbert has received material support (equipment) from Neuronetics. Dr. Elmaadawi receives research support from Duke University and University of North Western. He receives research fund from Neurocrine, Inc and Neuronetics. He is also a speaker for Neuronetics. Dr. Croarkin has received research support from the National Institute of Mental Health, Neuronetics, and NeoSync. He has received material support from and provided consultation to Myraid. He has consulted for Procter & Gamble Company. Dr. Demitrack is a consultant to Neuronetics, Inc. and a full-time employee of Trevena, Inc. All other authors have no competing financial interests.