Prevalence and Incidence

Lifetime prevalence is the proportion of people who have had a particular disease or disorder at any point in their lifetime. A recent large-scale survey of adults and adolescents in the United States concluded that the lifetime prevalence of AN is 0.9% among all women and 0.3% among female adolescents ages 13–18. ^{Footnote 3} Studies in other countries report lifetime prevalences of AN at up to 4.3% of women. ^{Footnote 4} The lifetime prevalence of AN among men is considerably lower than that among females. ^{Footnote 5} Studies typically underestimate the actual prevalence of AN, because patients tend to conceal their illness.

Although prevalence is an indicator of the impact of a disease on public health and the demand for care, incidence is a better indicator of patterns of disease onset. Incidence is the number of new cases in a population over a specific period. The incidence rate of AN is commonly expressed in terms of new diagnoses per 100,000 individuals per year. The reported incidence of AN ranges from 4.7 to 7.7 per 100,000 women per year. This statistic includes women of all age groups. But several incidence studies show that AN disproportionately impacts adolescents. ^{Footnote 6,Footnote 7} The peak age of onset of AN is between 13 and 18 years. ^{Footnote 8} Although onset of AN is rare in children, it does occur in children under 12 years old. ^{Footnote 9} The incidence statistics quoted derive from retrospective studies of healthcare records and so do not capture new cases of AN among patients who do not seek treatment. The already high—and rising—incidence of AN in adolescents and children demands attention to the etiology and impact of illness on adolescents specifically, as well as treatment protocols with their specific needs in mind.

Impact and Mortality

Even though the lifetime prevalence and incidence of AN are low relative to other mental illnesses ^{Footnote 10} and other eating disorders, ^{Footnote 11} the severity and impact of AN has been repeatedly demonstrated through studies of social impairment, medical complications, and comorbidity. The impact of AN on the developing adolescent is particularly severe. AN has profound psychological consequences that can impact adversely the developmental tasks of adolescence and young adulthood. Depression, anxiety, social withdrawal, difficulty eating in social situations, heightened self-consciousness, and fatigue associated with AN may isolate an individual from the tide of normal development. ^{Footnote 12} Serious growth retardation, pubertal delay or interruption, peak bone mass reduction, and abnormalities in brain structure can occur early in the course of the illness. ^{Footnote 13} Cardiovascular, gastrointestinal, and hormonal abnormalities are also common. ^{Footnote 14}

Most concerning, AN has the highest mortality rates among all mental illnesses and eating disorders. ^{Footnote 15,Footnote 16} High mortality rates are due not only to medical complications but also to suicide, the second most common cause of death among those with AN. ^{Footnote 17} The impact of AN and risk of death increase as the illness continues, suggesting a need for early treatment, as well as a need for specialized care as AN progresses to a more severe and long-standing stage.

Prospects for Recovery

All told, about 50% of AN patients make a full recovery with prolonged behavioral and/or medical treatment, 30% make only partial recoveries, and 20% do not recover, even with repeated attempts at treatment. ^{Footnote 18,Footnote 19} A substantial number of AN patients continue to display clinically significant symptoms for years and, in some cases, for decades. Adolescents are more likely to recover than those first diagnosed as adults, likely due to the inclusion of family therapy in adolescent treatment protocols. Adolescent onset is widely considered an indicator that a patient will have a positive outcome but is not unequivocally supported as a favorable prognostic factor in all studies. ^{Footnote 20} Those diagnosed before puberty, for example, typically have worse outcomes. ^{Footnote 21,Footnote 22} It is not known why some individuals recover and others do not, and no longitudinal data exist from which a clinician can gauge with confidence the outcome of treatment. ^{Footnote 23} Inevitably, many patients diagnosed with AN, including adults, adolescents, and children, will progress to long-standing AN.

There is no consensus definition or hard-and-fast rule for designating a patient as having long-standing AN, also called end-stage AN, chronic AN, or severe and enduring AN. I use the term “long-standing AN” (L-AN) because it does not imply that AN at this stage is untreatable, as “chronic” and “enduring” might imply. Some working definitions have been proposed and are used to guide clinical practice and research: entrenched patterns of AN behavior, enduring body mass index under 17.5, duration of illness greater than 10 years, extremely limited social life, or repeated treatment failures. ^{Footnote 24} To my knowledge, there are no available statistics on the demographics or epidemiology of L-AN specifically. This may be due to the lack of consensus over its definition, or poor follow-up at this stage of illness.

Unfortunately, there are few therapeutic options for L-AN. ^{Footnote 25} A conventional therapeutic regime for L-AN does not differ from the approach used during early stages of AN. It consists of refeeding to stabilize weight, followed by some form of therapy, sometimes with pharmaceutical management of depressive symptoms and delusions. An ongoing cycle of refeeding, followed by therapy for some time, and then relapse into significant weight loss, which necessitates further refeeding, is the norm for those who live with L-AN. Management of the L-AN patient may devolve into relatively unfocused, intermittent supportive interventions. Goals of care become unclear as patients grow more resistant to therapy. Clinician burnout and frustration are common among those who treat patients with L-AN. ^{Footnote 26}

Evidence Base for Standard Treatments

Recovery from AN is marked by normalization of the core symptom characteristics of anorexia nervosa, that is, involving weight, menstruation, and eating behavior. Standard treatment involves stabilization of medical symptoms, sometimes with refeeding, followed by behavioral interventions (i.e., cognitive behavioral therapy, cognitive analytic therapy, and nutrition programs) and/or medical management of symptoms (i.e., antidepressants, hormones, and nutritional supplements).

A recent comprehensive review of the evidence for AN treatment concluded that evidence for AN treatment is weak. Only 32 studies identified as addressing treatment efficacy for AN were published between 1980 and 2005, and there are serious shortfalls in the AN treatment literature. ^{Footnote 27} It is devoid of medical studies for adolescents, as drug trials have focused exclusively on adults; moreover, males are underrepresented in clinical trials of AN, and the majority of trials fail to report the race and ethnicity of participants. No clinical trials for AN address the optimal approach to inpatient weight restoration, the first step of any treatment protocol. This is especially concerning given the potentially lethal combination of fluid and electrolyte shifts that refeeding initiates. ^{Footnote 28} Most pertinent for the purposes of my argument, there has been only one controlled study of the treatment of long-standing AN in adults, ^{Footnote 29} and there have been no controlled trials on the treatment or management of L-AN in adolescents, despite the fact that some have argued for distinct treatment protocols at this stage of illness. ^{Footnote 30,Footnote 31} The prevailing need for effective methods of management, if not treatment, of L-AN is largely ignored in the clinical literature.

Ethically Salient Scientific Factors

Several of the scientific factors about AN are ethically salient to determining whether to enroll anorexic adolescents in clinical research on AN. AN is prevalent in children and adolescents and has a significantly higher incidence in adolescents than in any other age group. As we see diagnoses of AN in children and adolescents rise, we face a persistent lack of evidence to guide treatment decisions. This means that developing definitive treatment for AN in the adolescent population, especially for those with L-AN, is in the long-term interest of every adolescent AN patient, as well as their parents. ^{Footnote 32}

The literature on treatment for AN points to an unmet need among all AN patients, and among adolescents in particular. Clinicians routinely make choices for their adolescent patients that are not based on scientific evidence. Though clinicians know that family therapy is more efficacious than other therapeutic options in adolescents, not all adolescent patients have the stable home life necessary for its success, and studies have shown that it will not be efficacious in all cases even when done correctly. Finally, the fact that no studies exist focusing primarily on the treatment and management of L-AN indicates that there is a full 20% of the population of AN patients for whom no vetted therapeutic option exists. We can assume that among that 20% are adolescents diagnosed early, but for whom existing treatment failed.

The unmet need for evidence-based treatment of L-AN among adolescents provides a professional, if not moral, obligation on the part of researchers and clinicians to promote and conduct research on L-AN. ^{Footnote 33} To meet this unmet need, I propose three recommendations.

First, researchers should find consensus definitions of the various stages of AN and should use those definitions consistently across studies to allow accurate comparisons. Duration of illness greater than 10 years is used as an indication of L-AN because studies have shown that recovery is least likely after that point. Clinicians and researchers should not use the point of no return as a clinically significant marker, however, because it implies that no attempt to treat L-AN will be efficacious. Ongoing AN-associated behaviors and low body weight after a number of failed treatment attempts are more inclusive definitions of L-AN, as treatment-refractoriness does not exclude adolescents from a classification of L-AN. This is a positive result, as identification of adolescents with L-AN is crucial to foster research on adolescents at this stage of the disease, whose particular needs are not well understood, and certainly not met by the existing literature on AN treatment.

Second, researchers should separate adult and adolescent populations for research purposes. High incidence among adolescent females and males indicates that AN may be experienced differently among adolescents and adults. Clinicians’ experiences with adolescents corroborate a difference between adolescent and adult AN. ^{Footnote 34,Footnote 35} Treatment options need to be tailored specifically to adolescents’ psychological and developmental needs.

Finally, research should focus on patients identified as having L-AN, and particularly on L-AN in adolescents, for whom the long-term psychological and medical consequences are especially severe. Adolescents with L-AN likely require different treatment than adolescents in earlier stages of treatment and may require different treatment than adults with L-AN, given their unique developmental needs. Research in adolescents, as well as in adults, with L-AN is urgently needed to determine whether this is the case, and to develop treatments to mitigate the psychological, social, and medical impact of adolescent AN. Research on the efficacy of DBS presents an opportunity to meet this need.

DBS in Children for Other Indications

There is a precedent for using and studying DBS during childhood and adolescence. The FDA approved a humanitarian device exemption (HDE) for DBS of the global pallidus internus, an area on the motor cortex, for use in treating dystonia in children and adults in 2003. Dystonia is a movement disorder in which sustained muscle contractions cause twisting and repetitive movements or abnormal postures. Approval of the HDE authorized marketing of DBS for dystonia even though the effectiveness of the device for treating dystonia had not been demonstrated. In approving the HDE, the FDA determined that DBS does not pose an unreasonable or significant risk of illness or injury, and that the probable benefit to health outweighs the risk of injury or illness from its use, taking into account the probable risks and benefits of currently available devices or alternative forms of treatment. Most significantly, the FDA approved DBS for dystonia in all patients over 7 with treatment-refractory dystonia. ^{Footnote 37} At least 39 patients under 18 have undergone DBS for dystonia, including 21 patients under 12 years old. ^{Footnote 38}

Researchers confronted several problems when treating children with DBS for dystonia. They had to accommodate children’s small size, decreased muscle and soft tissue mass, and growth potential. Researchers now typically implant the battery-driven stimulator at the rectus abdominis to increase soft tissue coverage. Doing so reduces the battery pack’s prominence. ^{Footnote 39} (Stimulators are typically implanted at the collarbone in adults.) Despite increased incidence of postsurgical complications in children who have undergone DBS, researchers’ observations and accommodations suggest that DBS is a life-saving and beneficial option for pediatric patients with intractable dystonia. ^{Footnote 40}

Researchers have also studied DBS’s efficacy in treating Tourette syndrome in adolescents. Tourette syndrome is a chronic, childhood-onset neuropsychiatric disorder. It consists of motor tics and at least one vocal tic lasting longer than one year. It has a profound psychosocial and neurocognitive impact on sufferers. At least six adolescents have undergone DBS for Tourette syndrome. ^{Footnote 41} Adolescents were often lumped in with adults in publications of small case series of the outcomes of DBS in Tourette syndrome. Some authors suggest that DBS in younger patients may be associated with better tic control and functional outcomes. ^{Footnote 42} Others argue that tics that could result in permanent injury may justify DBS at an earlier age. ^{Footnote 43} Tourette syndrome–associated behaviors improved with DBS in most cases, and studies reported no serious adverse effects. DBS was not effective in managing symptoms in all cases, however, and one adolescent patient requested to remove the device. ^{Footnote 44}

DBS for AN

There are three reasons to believe that DBS might be an effective treatment for AN: (1) theoretical plausibility, (2) case reports from trials of DBS for OCD and depression, and (3) case series of DBS for AN.

Over the past two decades, breakthroughs in brain chemistry and neuroimaging have improved our understanding of the neurobiological bases of AN. Brain chemistry is disturbed in anorexic patients. Their levels of serotonin and leptin, which are associated with mood and appetite regulation, are abnormally high. Studies show differences in serotonin and dopamine receptor activity, indicating dysregulation of mood, anxiety, appetite, and impulse control. ^{Footnote 45,Footnote 46} In addition, brain scans show that some areas of the brains of anorexic individuals appear to be overactive, whereas others are underactive compared to controls. Anomalies are found in the cingulate gyrus, the nucleus accumbens, and the insula. All three of these midbrain structures are implicated in other psychiatric illnesses and signal dysregulation of emotions, anxiety, and reward processing. ^{Footnote 47,Footnote 48,Footnote 49} A neurobiological understanding of AN supports the hypothesis that DBS may be an effective treatment for AN because DBS aims to restore normal functioning to areas of the brain that function abnormally during illness.

Two case reports from trials of DBS for depression and OCD spurred interest in using DBS for treatment of AN. Mimi Israel et al. published a case report on Ms. A., who was treated with DBS for chronic depression. A. had suffered from AN since age 17. At age 52, A. underwent DBS. Four years later, in addition to recovering from depression, A. had maintained a healthy BMI. She had one relapse into AN following a friend’s suicide but otherwise required no further interventions for AN. ^{Footnote 50}

Shortly thereafter, Nicole McLaughlin et al. published a case report of a woman whose AN was markedly improved after undergoing DBS for OCD. The patient had childhood-onset OCD and AN. She underwent DBS at age 48. By the four-year follow-up, she had increased her food intake and variety. She attended events that involved eating and reported that she “doesn’t mind” going out to eat. ^{Footnote 51}

Publication of these case reports prompted further study of DBS specifically for AN, and a few case series were subsequently reported. The first case series, published by Jing Wang et al., involved only two young women. The study took place in Xi-an, China. The measure of success in this study was maintenance of a normal BMI, as well as improvement in symptoms of depression, anxiety, and obsessive-compulsive disorder; personality; intelligence; memory; quality of life; and social functioning. This study did not report on the average duration of illness prior to undergoing DBS, but both patients were over 18. Its authors concluded that DBS in the nucleus accumbens was an effective treatment for severe AN in young women. ^{Footnote 52}

Around the same time, a second study in Shanghai showed promising results in four adolescents. The average age of the patients before surgery was 16.5 years, and the average duration of illness was 19 months. Hemmings Wu et al. reported an average 65% increase in body weight at the 48-month follow-up. ^{Footnote 53} Despite promising results, the authors were criticized for having enrolled minors. Critics alleged that the researchers may have violated ethical principles by enrolling patients who had not exhausted treatment options or failed current treatment guidelines. ^{Footnote 54} Selecting subjects earlier in the progression of AN may have inflated results that may not be generalizable to adolescents with L-AN.

A group of researchers from Toronto published a case series of DBS for AN composed of six women with AN ranging from 24 to 57 years old. ^{Footnote 55} Their mean duration of illness was 18 years. Together, the study participants had been hospitalized almost 50 times. Nine months after starting DBS, three of the patients had sustained a body weight in the normal range. Two patients stayed more or less the same. One patient—the sole subject who did not suffer from a psychiatric disorder in addition to anorexia—deteriorated during the nine months after DBS began. The study methodology, inclusion criteria, and postsurgical consequences of DBS were much more rigorously described in the Toronto study than in the Shanghai and Xi-an studies, but the follow-up was shorter.

Risks Associated with DBS

DBS requires brain surgery, an inherently risky procedure. Placement of the DBS leads is associated with a 2%–3% risk of intracerebral hemorrhage and a 5%–8% risk of infection. There is a 0.6% risk of death related to surgery. ^{Footnote 56} Common adverse effects include seizure, infection, and pain. ^{Footnote 57}

Patients with AN would face increased medical risks from surgery; anesthesia is especially risky. Although only local anesthesia is used during lead placement, typically general anesthesia is used to surgically implant the battery-driven stimulator. AN causes a host of cardiovascular, respiratory, and endocrine abnormalities that complicate anesthetic management. ^{Footnote 58} Like children with dystonia, patients with AN have decreased muscle and soft tissue mass, and growth potential would have to be accommodated.

Other effects of DBS are not easily quantifiable, or even describable. DBS can have a profound effect on the lives, specifically the emotional lives, of those who have undergone the treatment for mental illness, for example, treatment-refractory depression. Several bioethicists and clinicians have written about the effects of DBS on quality of life, identity, and free will. ^{Footnote 59,Footnote 60,Footnote 61,Footnote 62} Some patients report feeling alienated from their desires and their self-conception post-DBS. In rare cases, patients have requested that stimulation be stopped and the leads and battery-powered stimulator removed following unforeseen psychological effects.

For adolescents who undergo DBS, the potential negative effects of DBS, whether psychological or biological, will last longer. Though DBS is reversible in that stimulation can be turned off and the device can be safely removed, it may produce intangible, lifelong psychological effects. These effects are not well understood and have never been tracked in adolescents. Beginning DBS during adolescence for treatment of AN will run the risk of adverse medical effects as well as of encountering unknown psychological outcomes. These risks are reasonable only when substantial benefit is expected from the trials and there is substantial risk associated with living with the disease under investigation.

The Study Participants

It is widely held that high-risk experimental interventions should be carried out only on patients who have exhausted other treatment modalities. Risky research should enroll those who have the most to gain and the least to lose. In these extreme cases, the risks of intervention are on par with the risks of continuing the status quo, and so the risk of undergoing a highly experimental intervention is deemed reasonable in the context. The same risks are not reasonable for a patient at an earlier disease stage, for whom there remain less risky treatment options to try. Within the population of people with AN, those with the most to gain and the least to lose are individuals who have already failed to improve after repeated treatment attempts, and so have progressed to L-AN.

Allowing adolescents to enroll in research on DBS carries risks and benefits for the study participants. Potential study participants face the aforementioned risks associated with surgery and unknown psychological sequelae. At the same time, enrolling in clinical research on DBS also provides benefits to study participants. Benefits include a chance at recovery and heightened surveillance for the duration of the trial, likely several years. If DBS is effective, study participants will obtain early access to treatment.

If adolescents with L-AN are not allowed to enroll in studies of DBS, they risk further organ system failure, cognitive decline, and psychological instability. The longer individuals progress with L-AN, the less likely it becomes that they will ever recover; adolescents who already have refractory AN are unlikely to lead healthy lives. Thus, although DBS for AN carries substantial risk, so, too, does living with L-AN. For young patients entrenched in the behaviors and complicated psychology of AN, DBS may be the best option for restoring normal life. That may be worth the risk of undergoing surgery and stimulation.

The Affected Population

Research into treatments for L-AN is urgently needed. The entire AN population suffers by not having good evidence available to guide clinicians in the treatment of their devastating disease, and the entire group benefits from studying the efficacy of newly available treatments. Moreover, DBS provides not only the opportunity to develop a new treatment for L-AN but also a chance to study the neurobiological underpinnings of AN. Regardless of whether DBS proves an efficacious treatment, its use may improve scientific understanding and lead to treatments down the line. The benefits of an improved understanding of AN will accrue to the affected population regardless of whether adolescents are included in trials of DBS for AN. It’s unclear, though, whether or how that knowledge would affect clinical practice for adolescents.

If adolescents are included trials of DBS, clinicians will know sooner rather than later whether DBS is efficacious in treating L-AN in adolescents, adults, neither, or both. Obtaining a positive answer sooner rather than later will prevent morbidity and mortality in adolescent patients, and obtaining a negative answer sooner rather than later opens up research time and effort for other projects. Moreover, even if DBS is not an effective treatment for L-AN in adolescents, patterns may emerge indicating that adolescents and adults require distinct treatment with DBS, improving scientific understanding of L-AN and its effects on developing and developed brains.

Current and future adolescents with AN and L-AN are harmed by not allowing teens to enroll. The status quo of management of L-AN will continue, and it will remain unknown whether DBS is an effective treatment for L-AN in adolescents. To exclude teens from promising research is to exclude other adolescent patients with L-AN from what may be the best treatment available in both the short and long run. Excluding teens from this promising line of research perpetuates the precedent of adolescent exclusion from research. Furthermore, if research continues on adults with L-AN only, and a negative result is found, it is less likely that researchers will explore its use in children. This would be an unfortunate result, as it may lead to missing a treatment that is effective in younger but not older patients. Adolescents may respond differently, so we should not predicate research on adolescents on data from adult trials in adults.

It is reasonable to conclude that the benefits to study participants and the entire affected population of enrolling willing teens in research outweigh the risks of undergoing DBS.