case study about anorexia nervosa

• Case report
• Open access
• Published: 15 February 2022

Terminal anorexia nervosa: three cases and proposed clinical characteristics

• Jennifer L. Gaudiani   ORCID: orcid.org/0000-0002-2035-9390 1 ,
• Alyssa Bogetz 2 &
• Joel Yager 2  

Journal of Eating Disorders volume  10 , Article number:  23 ( 2022 ) Cite this article

82k Accesses

35 Citations

453 Altmetric

Metrics details

Most individuals with eating disorders will either recover, settle into an unrecovered but self-defined acceptable quality of life, or continue to cycle from crisis to relative stability over time. However, a minority of those with severe and enduring eating disorders recognize after years of trying that recovery remains elusive, and further treatment seems both futile and harmful. No level of harm reduction proves achievable or adequately ameliorates their suffering. In this subgroup, many of those with anorexia nervosa will experience the medical consequences of malnutrition as their future cause of death. Whereas anyone who wishes to keep striving for recovery despite exhaustion and depletion should wholeheartedly be supported in doing so, some patients simply cannot continue to fight. They recognize that death from anorexia nervosa, while perhaps not welcome, will be inevitable. Unfortunately, these patients and their carers often receive minimal support from eating disorders health professionals who are conflicted about terminal care, and who are hampered and limited by the paucity of literature on end-of-life care for those with anorexia nervosa.

• Case presentation

Three case studies elucidate this condition. One patient was so passionate about this topic that she asked to be a posthumous co-author of this paper.

Conclusions

Consistent with literature on managing terminal illness, this article proposes clinical characteristics of patients who may be considered to have a terminal eating disorder: diagnosis of anorexia nervosa, older age (e.g. age over 30), previous participation in high quality care, and clear and consistent determination by a patient who possesses decision-making capacity that additional treatment would be futile, knowing their actions will result in death. By proposing the clinical characteristics of terminal anorexia nervosa, we hope to educate, inspire compassion, and help providers properly assess these patients and provide appropriate care. We hope that this proposal stimulates further expert consensus definitions and clinical guidelines for management of this population. In our view, these patients deserve the same attendant care and rights as all other patients with terminal illness, up to and including medical aid in dying in jurisdictions where such care is legal.

As a patient with severe and enduring anorexia nervosa advocating for my legal right to MAID (medical aid in dying), I confronted numerous obstacles and challenges from the medical profession, related not just to the question of whether I should have access to MAID generally, but more so, how my anorexia, a psychiatric condition frequently misunderstood by the medical community, interacted with my decision making capacity and desire to pursue MAID as one potential option knowing that my illness was indeed terminal. –Alyssa

The vast majority of potentially terminal illnesses carry with them thoughtfully considered and evidence-based staging criteria. These criteria allow patients and clinicians to distinguish mild and likely curable presentations of the disease from irreversible, pre-terminal and terminal stages. Medical specialties treating cancer, organ failure, or various infectious diseases have dedicated considerable attention and resources to delineating levels of severity. While preliminary suggestions for labeling severe and enduring anorexia nervosa (SE-AN) [ 1 ] and for staging the disorder have been proposed [ 2 ], generally accepted staging criteria for anorexia nervosa (AN) have not yet been developed. Remarkably, the same diagnostic label (“AN”) and treatment criteria that apply to teenagers only a few months into their disorder are also used for patients who are decades older, who have lived through innumerable admissions to inpatient and residential care facilities, and whose quality of life has been irrevocably damaged by persistent, severe mental and physical illness. The field acknowledges SE-AN as a somewhat distinct clinical condition, but despite thoughtful clinical and research efforts [ 3 , 4 ] the designation has not been formalized as a diagnosis, and consensus regarding criteria for SE-AN remains elusive [ 5 ].

AN carries the second highest mortality rate in the DSM-5 after opioid use disorder, with a death rate estimated at 5–16 times that of the general population. [ 6 , 7 ] Several important recent studies confirm and expand upon these data. A specialized medical inpatient unit in France for those with severe anorexia nervosa evaluated 384 patients admitted over 17 years, with a mean age at admission of 29.4 years old. The standardized mortality ratio (SMR) was 15.9 for women and 22.4 for men, where older age was determined to be a major predictor of mortality. The mean age at death was 41.3 (± 15.3) years, on average two years after hospital admission. The SMR was maximally increased for patients whose first admission to the unit took place while they were between 25 and 35 years old. Specifically, those admitted between 30–34 years old had the highest SMR of 26. Somatic (medical) causes accounted for 43% of deaths, while 11.5% of deaths were caused by suicide. [ 6 ] In a registry-based observational epidemiological study encompassing the entire population of Denmark over 44 years, the SMR for all-cause mortality reached a maximum of about 6 in the age group 20–34 years, and the SMR for suicide in those with AN was 11. Natural causes accounted for two-thirds of death in those with AN. [ 8 ] Finally, a retrospective cohort study evaluated 19,041 individuals with an eating disorder in Ontario, Canada, using administrative healthcare data. The entire cohort, not comprised only of those with AN, had an SMR of 5; they found that potential years of life lost were 6 times higher than expected compared with the Ontario population. Similar to the other studies, peak values for SMRs were observed among adults between 30 and 44 years old, and again the SMRs observed in males were almost two-fold higher than in females. [ 9 ] Importantly, the profound suffering inherent in AN drives the high suicide rate noted in multiple studies, where up to 20% of patients who die prematurely do so by suicide [ 10 ]. Compared with gender- and age-matched groups, patients with AN are 18 times more likely to die by suicide [ 11 ].

Based on these data, AN can unquestionably prove fatal. Despite this fact, the field lacks clinical roadmaps for compassionate, appropriate care for those who will not be able to survive. This does great disservice to patients and their families. By comparison, we do not expect individuals with metastatic lung cancer who have disease progression despite past treatments, which often come with negative sequelae, to keep presenting for those same ineffective treatments. Rather, they are more likely to receive the psychological preparation, connection, and medical and emotional support offered to patients with terminal conditions. Although current laboratory measures and imaging studies by themselves are unable to help us stage patients with AN, based primarily on clinical histories and patients’ narratives we can better understand the clinical course of this illness and the subset of patients with AN who may seek palliative care [ 12 , 13 , 14 ].

There is growing recognition that palliative care may be appropriate for some patients, but the clinical characteristics for terminal anorexia nervosa have not been proposed. Delineating and validating this stage would greatly assist patients, families, and clinicians across disciplines, especially those in palliative and hospice care. Designating terminal AN may more readily enable patients to receive palliative care, hospice care, and emotional and practical resources for loved ones, as well as access to medical aid in dying (MAID) where legal. Therapeutic goals in these situations are to ameliorate suffering and honor the life lived. Of note, MAID is offered to individuals whose death is inevitable within six months from an underlying disease process; it provides patients a choice in how they die, not whether they die. It is not a means of suicide.

In this paper, we describe three cases of exceptional people whose AN was terminal, and who died peacefully with family around them. All three were patients of a private practice outpatient medical clinic specializing in eating disorders in Denver, Colorado. One patient, who had been a medical researcher herself, was so passionate about the topic that she asked to join as a posthumous author on this paper so her voice could be heard. The other two patients’ parents consented to share their son’s/daughter’s stories and reviewed, and all three families edited the relevant story prior to manuscript submission. All families agreed that first names should be used instead of a pseudonym or initials in order to emphasize the truly personal, real-life origin of these stories. Based on these experiences and others [ 12 , 14 ], we conclude by proposing a set of clinical characteristics of those who can be identified as having terminal AN.

Case presentations

Case 1: aaron.

Aaron was a 33-year-old man with a long history of restrictive AN, severe obsessive compulsive disorder (OCD), recurrent major depression, and chronic suicidality. He had been a sensitive child with low self-esteem and perfectionism from a young age. His parents noticed OCD traits from early childhood, but he did not receive this formal diagnosis until years later.

During his freshman year in high school, a health class warned about the risks of “eating junk food.” Aaron began to run regularly and played hours of basketball daily. One by one, he eliminated dietary fats and created food rules. His parents thought this was just a stage, an assessment they came to understand very differently over time, but they eventually recognized his serious problems and established a treatment team. Later in high school, Aaron was hospitalized several times for AN, participated in family therapy, and required his mother’s presence, even at school, to complete meals. Despite graduating as valedictorian of his high school class, he was initially too ill to start college. Later, his attempt to begin college was thwarted by his need for constant supervision of food intake. Aaron’s perfectionism and self-criticism ultimately ended his college career.

Over the next two decades, Aaron spent countless months in medical hospitals and in inpatient and residential eating disorder settings. He repeatedly gained the weight required for discharge so that he could return home, only to inevitably relapse. He felt mortified and guilty about the amount of money his family spent on his treatment, and he was acutely aware of life passing him by.

In his early 30 s, following a long and serious downward spiral during which he refused a higher level of care, his family finally threatened to call 911 if he did not enter treatment. Consequently, he was admitted to an inpatient eating disorder program. After first spending time in a hospital setting for stabilization where he refused to eat, a feeding tube was placed. A court mandate to ensure ongoing treatment was requested and granted on grounds of grave disability from his mental illness, and he spent the next 10 months against his will in inpatient and residential eating disorder care. He cut off communication with his parents but allowed the treatment team to talk with them. Eventually his therapist convinced him to have family sessions over the phone; it was the only time his parents could talk to him.

Aaron’s persistent resistance to treatment throughout his stay at the eating disorder program caused difficulty in maintaining his nutritional stability. He underwent in-depth exposure and response prevention therapy around food as he continued to be tube fed, and he was finally able to sustain his weight with oral food. At that time, he only agreed to eat to avoid the prospect of being administered olanzapine against his will, as he feared this medication would cause him to gain weight.

After Aaron had been fully weight restored for several months, he was stepped down to a partial hospital program (PHP). He immediately restricted intake and proceeded to lose nearly a pound a day, resulting in readmission to residential treatment where, after intensive efforts, he once again achieved his target weight. Aaron struggled with basic activities of daily living due to his OCD. For instance, he resisted using lotion or lip balm as he feared they might be absorbed into his skin as calories. He completed a course of intranasal ketamine in hopes of alleviating his OCD, depression, and suicidality, but ketamine treatments had no meaningful impact.

Author JG (hereafter referred to as “Dr. G”) first met Aaron for an outpatient medical consultation after he had completed this most recent residential treatment and was once again about to step down to PHP. This consultation constituted one component of an organized, comprehensive future discharge plan. In this initial medical visit, after a year of residential treatment, Aaron was medically stable. He desperately missed his eating disorder behaviors, fantasized about eating less and losing weight, and wished his AN would have already taken his life. Aaron mused that his all-or-nothing, perfectionistic temperament made the unknown terrifying, but he felt proud of how rigidly he had previously adhered to his eating disorder rituals, as he believed that few others could achieve a similar degree of calorie restriction. Despite his long history of treatments, including his year of suffering through the long court-mandated treatment, Aaron had never meaningfully changed his eating related attitudes, thoughts, or behaviors. He had absolutely no motivation for recovery.

During the initial consultation, Dr. G informed both Aaron and his emotionally supportive and highly invested parents that she could offer ongoing outpatient medical care along one of two pathways. In one, Aaron would complete PHP, be discharged to home (where he lived with his parents), see a therapist and dietitian regularly, and work on whatever degree of recovery he could bear, aiming toward a quality of life that he called “living productively.” Should he decide that he required a higher level of care, the team would promptly support that choice. On the second pathway, if Aaron relapsed and declined readmission, the outpatient team would no longer battle with him to seek a higher level of care, given the futility of his most recent, autonomy-depriving treatment course. Rather, the family and team would support and comfort him until such time as he required home palliative care and eventually hospice support. Aaron initially felt that these two choices were needlessly stark and binary, and he settled back into PHP.

Two months after the initial consultation, Aaron continued to endure PHP, primarily to honor his commitment to his residential treatment team that he would see through his course of treatment. But he felt no better. He received a course of intravenous ketamine to supplement the intranasal ketamine treatment initiated in the residential program, but he experienced no improvements in mood, hopelessness, or OCD. Just before discharge to his parents’ home, he still hadn’t decided which treatment pathway to choose. He would not accept psychiatric medications, and a team consisting of a physician, therapist, and dietitian was established to care for him at home.

However, starting on the day of his discharge from PHP, Aaron stopped eating altogether, a course of behavior that is rare even in those with severe AN. He drank only water, stating, “I don’t want to die, but my eating disorder is in charge.” A week later, he met with his longstanding outpatient therapist. She was very apprehensive about his ability to remain in the community, and she felt ambivalent concerning what her role might be if he insisted on remaining at home. Aaron told her, “I wish I could eat, but I won't eat; I don't want to die, but I feel hopeless that there's any other pathway.” The therapist worried that she would be forced to have Aaron detained against his will in his home state. A formal decision-making evaluation was performed by a local psychiatrist, and Aaron was found to possess decisional capacity. Consulting with his home therapist, Dr. G proposed that Aaron’s refusal to eat was less about “wanting to die” than simply accepting that he could not live—he was not “attracted to life” [ 17 ]. Dr. G suggested that the proper course at this point would be to proceed with a home palliative care consultation and shift treatment goals to supporting comfort and dignity, as Aaron clearly declined a return to treatment.

During a telemedicine meeting with Dr. G a week later, Aaron asserted, “I don't want to do this for anybody else anymore. It's time to do things only if I want them.” At about this time, Aaron also sought comfort from his therapist and his religious leader, as the prospect of death frightened him, and he was unsure what dying would mean. But he described that being given the choice of what would happen next was empowering – “different, scary, relieving, and right” – in great contrast to repeatedly feeling powerless and demeaned by his many prior chaotic relapses followed by intense pressures to return to treatment. Aaron signed a Do Not Resuscitate (DNR) order and within the next few days was referred to a home palliative and hospice care organization. Dr. G spoke with the organization’s medical director to explain why this brilliant 33-year-old man who was refusing to eat was being referred for palliative care. Aaron hoped that the home palliative care service would help him and his family “process this sorrow and fear.”

Two weeks later, after more than a month of eating nothing and drinking only water, Aaron’s OCD and insomnia were heightened; he worried that by simply smelling his mother’s cooking he might be ingesting those calories. He believed that he might absorb calories from the grocery cart of the person ahead of him at the store. Always reluctant to take medications, he began to consider accepting anxiolytics from the hospice staff, whom he thought were extremely kind.

Aaron noted that by spending no energy forcing himself to eat, he was able to direct energy toward engaging in his faith. His siblings came to visit, and as they talked and laughed, he realized it had been years since they had connected positively. Throughout the course of his eating disorder, every family connection had felt fraught. He summarized his collective family’s response to this pre-terminal phase as, “They were very supportive. They recognize the gravity of this situation. They aren't angry, sad but not fearful.” As he chose to spend his days talking with his parents and sleeping, he noted that he was thinking about others "rather than being so self-absorbed." Imagining his parents’ distress made him sad, and he wanted his parents to keep getting support after his death. "This is one of the hardest things they've had to deal with in their lives."

Even as he rapidly lost weight, Aaron’s body distortions grew worse, and he kept wishing his weight would fall even faster. After almost six weeks without food, Aaron began accepting anxiolytics and antiemetics. He obsessed that someone might have injected his water bottles with calories. When Dr. G asked if he had any words to share for posterity, he expressed words of warning for those who might find themselves in his situation: “OCD will amplify,” “Be prepared for an annoying obsessive brain that might drive you crazy,” and “Just because you aren’t eating doesn’t mean it’s all good now.” He expressed how vital it was to “have people in your life [doctor, parents, family, close friends] whom you trust and can seek reassurance from, who love you unconditionally,” whose comforting words can be “life-saving in terms of giving you peace.” He connected deeply with a feeling that peace comes from God.

After about eight and a half weeks without any food, Aaron was spontaneously vomiting daily and feeling much weaker. Beautifully cared for by home hospice, he began to take low dose morphine for pain and distress. When Aaron’s parents wondered what his death certificate would say about cause of death, Dr. G reassured them that the cause would be anorexia nervosa and malnutrition, not suicide. Often speaking through tears, Aaron’s parents described how they were enjoying a deep loving sweetness with their son that they hadn’t experienced in years, and how they would miss him when he died. They felt compassion for those who lose a loved one abruptly without having time for love, connection, and closure. Often, they saw glimpses of the boy they hadn’t seen in years, as when he looked at photos that made him laugh.

Two weeks later, Aaron passed away with his family surrounding him. Even as they were exhausted and grieving deeply, his parents expressed enormous gratitude for the care he received and for the way they had been able to reconnect with him.

Case 2: Jessica

Jessica was a 36-year-old woman with a history of OCD and AN, purging subtype (laxatives) that began during her junior year of high school, when she tried to lose weight prior to a vacation. This started a pattern of restricting, binge eating, and then overexercising that persisted into college. When her weight, which had remained normal for some time, did eventually drop, she left college for intensive outpatient eating disorder treatment. It was such a difficult experience that from this time on, she mistrusted eating disorders providers. She lamented that she lost most of the fun of college to her eating disorder.

Due to progressive constipation, Jessica began using laxatives, which led to laxative abuse. She soon found that every time she stopped taking laxatives, her weight skyrocketed (due to rehydration and rebound edema). Ultimately, her AN caused her to drop out of nursing school. Jessica experienced her first hip fracture from severe osteoporosis when she was critically emaciated at age 27, requiring her to move home with her parents; the following month she incurred a stress fracture of her shoulder from using crutches. Her parents pursued guardianship as Jessica was refusing a higher level of care, but her medical team refused to release records to the family’s attorney due to HIPAA. Without the option to pursue guardianship with mandated longer-term residential treatment, her parents came to believe this was the critical juncture where recovery might have been possible, but instead her disease became more entrenched. Over the next 7 months, working with her outpatient primary care provider, dietitian, and therapist, she slowly gained a meaningful amount of weight, although she remained very underweight. Following this, she got an excellent job and once again lived independently from her parents, working productively for three years. However, at age 29, her increased anxiety, the side effects of laxative abuse, and the shame of her anorexia caused her to separate herself from her family and to work from home, increasing her isolation. She checked herself into an expert inpatient medical center to stop using laxatives and then spent a week in inpatient eating disorder treatment before leaving against medical advice. Jessica did manage to stay off laxatives for a year but was plagued by edema. Repeatedly, restriction and overexercise would recur, usually accompanied by laxative abuse, which at its worst consisted of taking 100 tablets a day.

During her initial consultation with Dr. G, Jessica memorably stated, “The eating disorder keeps me out of integrity with my values. It doesn't feel good. You believe something but aren't living it. This is the biggest motivation for wanting to change. I really want to live in alignment with my values, honoring my body, feeling things, stopping being unkind to my body.” Although very kind and compassionate towards others, she struggled to show herself the same grace.

Jessica met criteria for immediate admission back to the inpatient medical service, but given her prior negative experiences with treatment, she wanted to attempt to keep working and live near her parents. She agreed to outpatient care with a multidisciplinary team, focusing on harm-reduction goals. Initially, Jessica was able to follow medical and nutritional recommendations faithfully. Then, within three months of initiating outpatient medical care she fell and sustained a pelvic fracture. This was frightening, disabling, and prevented her from taking her calming (and to her, calorie-burning) nature walks. Overcome by managing the challenges of a rapidly changing body on her own and worried about her fracture and bone health, Jessica readmitted herself to specialized inpatient medical care for medical stabilization. Following stabilization, she agreed to transfer to residential care to attempt a full course of eating disorder treatment. However, after two weeks in the residential care program she left against clinical advice, unable to follow the meal plan consistently and feeling extremely distressed by her bodily changes (even though her weight had barely changed).

At home, Jessica again tried hard to follow treatment recommendations at a harm-reduction level (no laxatives, low caloric intake, gentle movement in the outdoors), but once more the distress of bodily changes was too much for her to bear. About a month after leaving the residential program, Jessica first talked about the possibility of palliative care and began talking with her mom about suicidal thoughts. Most nights she would say she hoped she didn’t wake up the next morning. In order to help Jessica resist the laxatives that gave her such severe abdominal pain and nausea, and still hoping to support her in finding an acceptable degree of harm reduction, Dr. G worked with Jessica to use diuretics to manage fluid weight changes. (Notably, this approach would rarely if ever be offered in a more typical eating disorders treatment plan.) Jessica operated within these guidelines and constraints for the next five months, at times thinking she might be able to persist, but more often lamenting that this strategy was still too difficult and painful. By this point, she had been granted indefinite leave from work and moved in with her parents.

About nine months after initial consultation, Jessica acknowledged that it was time for a palliative approach, confessing, “I’m just ready. It's been a long fight. I'm eating so little, and I'm back on the laxatives every couple of days.” She declined intranasal or intravenous ketamine which might have ameliorated her depression, OCD symptoms, and hopelessness. As she felt progressively miserable physically and psychologically, her suicidality increased. She purchased a gun, and one night she drove to a bridge with thoughts of jumping off, but then decided to return home. She had difficulty finding a therapist who understood terminal AN and who could accept her treatment trajectory, but she found and worked with a kind naturopathic doctor who specialized in mental health, and she did experience some benefit from psychiatric medications.

At this point, fearful of suffering a long, drawn-out death from starvation and unwilling to put her parents through the agony of witnessing this decline, Jessica requested referral to a palliative care specialist who assessed patients for medical aid in dying (MAID). Dr. G spoke with Jessica’s parents repeatedly, assuring them that guardianship and forced treatment were likely now to be futile. The parents had done everything possible to help their daughter find an acceptable quality of life. Jessica signed a DNR order. After speaking with the palliative care physician by phone to discuss the case and advocate for Jessica, Dr. G completed the MAID forms as consulting physician, given that Jessica’s prognosis was presumed to be 6 months or less. The palliative care physician prescribed the MAID medications.

About a year after the initial consultation, and about three months after the MAID consultation, Dr. G saw Jessica for the last time via telemedicine. Jessica wrote to Dr. G in an e-mail, “I’ve been back in a place the last several weeks where the emotional pain and the physical and emotional exhaustion of living like this are just too much for me. I’m trying to make it to the end of May, maybe through June to meet my brother’s upcoming baby before I go.” Jessica described her life as filled with unbearable pain and anxiety. Watching people walking around the neighborhood making future plans felt devastating, because she’d “give anything to be in anybody else’s shoes.” Yet when she thought about stopping diuretics, eating enough food, and gaining weight so she could physically live that life, she said, “it feels impossible.”

Jessica waited several weeks to fill the MAID prescription. She then set multiple dates to use it over a couple of months and changed her mind as that date got closer. A month before her death, she started to receive home hospice services. During this time period, she had long conversations with her parents, brother, and friends, noting that she had many happy memories over her life, apologizing for what she had put them all through over the years, and stating that she hated her eating disorder. She told them she realized that, while her eating disorder behaviors made it seem like she hadn’t loved or trusted them at times, she loved them all very much. She repeatedly told her family that she didn’t want to die, that she didn’t want to miss out on future time with her family, friends, and niece and nephew, but she just couldn’t continue to exist this way. The emotional pain and anxiety were unbearable. She couldn’t live a normal life, and she felt her body was too destroyed to recover. Her parents believe that in her last month she was trying to die naturally by barely eating, reducing her fluid intake, and walking for hours daily, even when she had to sit down often to catch her breath. She stopped driving and carried identification in case she collapsed on a walk. She fainted at home several times in the week before her death, including the night before she died. On the day she took the MAID prescription, she stayed in bed, was at peace, and spent time talking with each parent and her brother. Together as a family, they reminisced, laughed, cried, had their “hug circle” as they had called it since her childhood, and felt surrounded by love. Her parents each held a hand, and her brother sat right next to her. During the three doses of the medicine taken over an hour, she was comfortable and conscious. Within ten minutes of taking the final dose, Jessica closed her eyes, and her breathing slowed.

Jessica didn’t choose to live with anorexia. For all the years she endured living within its prison and myriad complications, her parents ultimately felt strongly that she deserved to choose the time, place, and way of her release. They felt that an unexpected blessing of MAID was that it allowed Jessica to live several months longer than she otherwise would have. Knowing she didn’t have to die a violent death by suicide, that she would have a peaceful way out when the pain and anxiety became unbearable, and that she would be able to die with dignity surrounded by loving family, allowed her to hold on longer. As a dying wish to her mother, she shared, “Mom, I'd like you to do something that will help others not go through what I went through."

Case 3: Alyssa

Alyssa, the posthumous author on this paper, was a 36-year-old woman with OCD, depression, and restrictive AN who described herself as having “a type A, neurotic personality: a sensitive, compassionate, loving person who's incredibly self-critical and has wanted to do things 0% or 110% with no gray area.” She first felt suicidal at age 13, when she realized that her body was too large to fit into standard dress sizes for her upcoming Bat Mitzvah. She started therapy at that time and was continually in therapy thereafter. After going through high school at a higher weight, the summer before college she vowed to change her body and began exercising in earnest. In college it was easy to restrict. By the time she returned home for Thanksgiving she had lost a substantial amount of weight. Everyone praised her, and she experienced “a deluge of external validation that was irresistible,” firmly establishing her eating disorder by age 18. Alyssa wrestled with AN throughout the rest of her education and career. A brilliant academic, she became the only non-physician Assistant Director of a major academic medical center residency department, mentoring residents and students, doing research, and publishing in major journals.

After struggling with AN for 15 years, during which she received intermittent outpatient support, Alyssa moved in with her parents and reduced her workload. She was extremely helpful in her mother’s struggle with a cancer diagnosis and often underplayed the significance of her own illness. At age 33, to correct severe hypercalcemia she was admitted to the teaching hospital in which she had previously worked. The family felt that her AN was hardly addressed during that hospitalization, in part due to the fact that institutional expertise for AN was confined to a pediatric program. To them, this felt like a vital missed opportunity to attempt changing her disease trajectory, in particular as the only recommendation on discharge was to seek residential eating disorder care.

Alyssa worked for 7 months to obtain insurance authorization for care in a residential eating disorders program, and to gain enough weight to meet their admission criteria. However, upon admission to that program she was deemed still too underweight (by one pound) and was referred to a specialized inpatient medical program. Being rejected for care after so much work also felt like a missed therapeutic opportunity. After a delay, Alyssa spent several weeks in the specialized hospital program and met the minimal criteria for discharge, departing with the understanding that she would immediately enroll in another residential program. However, after discharge from the hospital she refused to do so and could never accept going to an eating disorders program thereafter.

In the years prior to initial consultation with Dr. G, Alyssa’s outpatient treatment team included a local primary care physician with whom she was very close, a therapist she had been seeing regularly in recent years, and an expert eating disorders therapist who had worked with her and the family over the years. Over a period of three years, Alyssa had intermittently thought about and even phoned Dr. G’s outpatient medical clinic, but she never booked an appointment, indicating that she felt very ambivalent about recovery and was considering a palliative care approach. When she finally presented for an initial consultation, Alyssa identified her goals as follows: “I really want a life, to use my Masters in Social Work degree to help others heal, to find a partner, and to experience pleasure, laughter, joy, and freedom, including from my own brain.” As her main barrier she cited the chronic, longstanding shame and body disgust that persistently kept her from meeting her own needs.

At the time of initial consultation, Alyssa met criteria for inpatient medical hospitalization, although she experienced remarkably few physical symptoms, which reinforced her view that she must be “fine.” She declined a higher level of care. Nonetheless, she saw herself as shamefully thin, more keenly felt given her extended family’s experience of the Holocaust. She wanted to be able to walk down the street without turning heads due to being so emaciated, but concurrently struggled to balance this desire against her strong resistance to gaining weight.

Alyssa agreed to ongoing care with the clinic and accepted referral to an expert registered dietitian. She committed to at least attempt a harm reduction approach in which she would slowly restore weight to a point where she could be more physically, mentally, and professionally functional, and where she could resume her yoga practice. However, she stipulated that she would halt weight restoration if and when her AN thinking could no longer bear it. Over the course of the next year or so, she valiantly succeeded in increasing her caloric intake considerably above her previous severely restrictive baseline. But due to the hypermetabolic state often seen in malnourished patients who increase their caloric intake, she experienced no meaningful weight gain.

Nine months after initial consultation, Alyssa emphatically reflected that her goals had not changed, but she had grave doubts about her ability to achieve them. She described feeling “utterly exhausted” and could no longer muster the strength to keep fighting. She vividly described her daily internal battles, struggling every minute of the day to eat enough of her meal plan and constantly fighting against the extreme headwinds of her AN’s resistance. Once she had eaten, she would bitterly berate and punish herself for having done so. At this point she was not certain that her AN was terminal, but she was moving strongly in that direction and wanted to understand her options.

Dr. G clarified that at any time, Alyssa could choose to pursue full recovery and a higher level of care, could continue fighting as she was, or could consider two options that did not focus on recovery. The first option would be choosing palliative care. This would acknowledge that she would likely not survive and also allow her to consider a "bucket list" of experiences for the time she had left. Palliative care would mean that she could eat what appealed to her, with no pressure applied by the team. The treatment focus would be on finding joy and comfort as much as possible. Dr. G emphasized the value of signing a DNR document to protect Alyssa from the mandates of the healthcare system in the event that she experienced an abrupt decline and/or cardiac arrest. Alyssa was also advised that a home palliative care/hospice evaluation would be useful to oversee her treatment as desired during this stage, for emotional and practical support if needed and to protect her parents from any potential legal repercussions should she pass away at home as an emaciated adult. Dr. G noted that for some patients, this stage can last a long time, and that some can “reset” when pressures to gain weight and threats of mandated treatment are removed. In some cases, this state of reduced external pressure might even lead to renewed ability to engage in meaningful harm reduction and even recovery work.

The second option would be to seek hospice care. Hospice care would be suitable if the torments of her AN and the extraordinary difficulties of moving about the world in a skeletal body were beyond being helped by a palliative care approach. Given her faster metabolism, if Alyssa abandoned her attempts to consume a higher meal plan, she would clearly have a less than six-month prognosis and qualify for hospice care. With this option, Dr. G would refer Alyssa to a home hospice service, anticipating that she would become increasingly frail. The home hospice staff would establish warm relationships with Alyssa and her parents, make sure that anxiety, insomnia, nausea, and/or pain were managed, and provide them all access to psychological and spiritual support as desired. During this time, Alyssa could live her life as she chose. As she became less independent, hospice would provide assistive aids such as a shower chair, bedside commode, and hospital bed. The overall goals would be to maximize Alyssa’s comfort, dignity, and time to connect with family.

During this conversation, Dr. G also noted that Alyssa lived in a location where MAID was legal. If she chose the hospice route—and had interest—a referral for the option of MAID was also possible. Alyssa was informed that she herself would have to administer the MAID medications if she chose to use them; no one else could administer them to her. After completing the required regulatory processes and filling the prescription, MAID medications could be used or not as desired. But, as the human body can be exceptionally resilient even with terminal malnutrition, having the medications at hand would give Alyssa the opportunity, while still having an intact brain, to choose not to suffer through additional weeks of extreme physical discomfort and weakness.

A week after these options were reviewed, Alyssa wrote Dr. G:

After deep reflection and discussion with my parents, I’ve decided it makes sense to initiate the Hospice process (Ie evaluation, etc.) now so my family and I are prepared for what may come. I would value your guidance and help with this….I do not know if they have ever worked with patients like myself… I would love for you to be the PCP overseeing this process regardless of the Hospice we select if, and only if, you are comfortable with this. I want to be clear that my priority is to obtain access to the medications that would support my legal right to die should I wind up choosing this path in the future. I feel strongly that based on our thorough discussion, I am aware of my options and their risks and benefits in light of the trajectory of my illness. Please do let me know what I can do to help facilitate initiation of this process. I am available and happy to help.

In a family meeting the following week, Alyssa’s father, a physician, tearfully shared the principles he and Alyssa’s mother had come to accept during intense conversations with their daughter: She had the right to choose care or no care after having been ill for 18 years. There would be no ultimatums. This disease would probably be the reason that "we lose you." They knew how much she had suffered and continued to suffer, and they understood that at some point the psychological anguish would become unbearable for her. They respected that this could be as bad as physical pain. They accepted that when the anguish became unbearable, Alyssa would have the right to end her life by taking medical aid in dying medications. They agreed that financial planning and end of life planning were worthy tasks. To Dr. G and to Alyssa, these words conveyed deeply reassuring love, compassion, and support.

Alyssa’s parents asked whether any treatments remained that might yet change the outcome of her course, specifically noting that Alyssa had not completed a full residential eating disorder program, never fully restored weight, never tried newer psychedelic options such as ketamine, psilocybin, or MDMA, and hadn’t had a feeding tube. Dr. G acknowledged that all but the feeding tube might ordinarily be undertaken prior to someone’s seeking end of life care for AN. Yet, she had been suffering for so long, and despite many conversations about all these treatment possibilities, Alyssa would not consent to any of them. Therefore, given her clarity of understanding around these issues and her sense that she could not fight anymore, everyone had to accept that they weren’t meaningful options. With regards to a surgical feeding tube in the context of AN rather than due an anatomical impediment, Dr. G noted that if someone restricts the “tube God gave them,” i.e. their esophagus, they would also be very likely to restrict through a surgical feeding tube, so that would not be a long term solution.

An excellent home hospice agency agreed to work with Alyssa and her family, and Dr. G placed a referral for a MAID consultation. The palliative care physician met with Alyssa about MAID. Since the idea of requesting MAID for a patient with AN was so foreign and unnerving to him, he asked Alyssa to be assessed formally for decision-making capacity. After a local psychiatrist confirmed that Alyssa clearly possessed decision-making capacity, the palliative care doctor fully accepted Alyssa’s right to enter home hospice care and could understand the rationale for MAID provision. However, even as he and his team provided empathetic support, he ultimately felt personally unable to write the MAID medication prescription due to his discomfort with the unique presentation. Clarification with the state’s Medical Board and other regulatory entities determined that Dr. G, licensed in this state although based in another state, could serve as prescribing physician, and that Alyssa’s longstanding primary care physician could serve as consulting physician. Dr. G prescribed the MAID medications about six weeks after Alyssa entered hospice care. Four days before her death, eager to contribute to this article, Alyssa sent Dr. G the following (unedited) notes about her thoughts on this complicated topic:

Below I share the considerations I made as I weighed the potential benefits and risks of pursing MAID. I share my experience in hopes of offering a first-hand perspective that may help other patients and physicians as they consider and weigh the option of utilizing MAID, rather than offering a prescriptive decision-making tool or recommending that all patients with terminal SEAN have access to such medication.

Personal considerations:

MAID not pursued in isolation, but rather in the context of being in Hospice care following a terminal dx of anorexia (i.e., estimated 6 months or left to live). I would not have qualified for Hospice care unless my illness was terminal (i.e., not reversible for me in light of physical, mental, emotional damage to my body).

In my individual case, death was inevitable. I clearly understood my prognosis and accepted this. I saw MAID as an opportunity to select a specified time and circumstances for my death. Death itself is fraught with fear, ambiguity, a sense of powerlessness and tremendous anguish, not just for the patient who is dying, but for that patient’s family. Upon deep reflection, I came to see MAID as an opportunity to relieve my suffering and minimize at least some of my family’s suffering related to my death by choosing the when and how of my death, rather than “wait” for sudden death from cardiac arrest or other outcome of my illness or experience a slow and protracted death as my family and I watch my body and mind degrade over days and maybe even weeks of time

I had to ask important questions about my quality of life and whether for me, the quality of my life was more important than the quantity of days I remained alive. I was experiencing extreme physical pain, was unable to walk, could not sit without discomfort, I couldn’t swallow my food, my breath was labored, and I had frequent chest pain. I was not living. I felt like “dead girl barely walking.” For me personally, a longer life spent in bed feeling ill and suffering and dependent on others to provide most of my care was not how I wanted to live. My concerns about this suffering trumped any fear of selecting the route of my death (again, knowing that death was inevitable). Knowing that I could utilize MAID if the suffering became so severe offered me a sense of ease and peace of mind in my final stage of life that I would not have had otherwise

One question that I needed to answer for myself honestly was whether I understood the impact use of MAID would have on my family. I had to confront that my use of MAID would be difficult for them, not just the idea of my using it but how their presence at the end of my death, watching me administer my own medications to die, would be ingrained in their memories of me as their daughter and their sister, and how this story of my passing would affect my family throughout the generations to come (i.e., what stories would they tell about my life and death, how could this be traumatizing or perhaps seen as healing?). Such questions could only be answered through ongoing involvement and discussion with my family members, which we had with my physicians and amongst ourselves

Another important question I asked was how would I want a family member to die if I knew their illness was terminal and death was imminent. Would I see their use of MAID as a compassionate act towards themselves? How would I tell their story? Would I extend the compassion I was asking for from them to them if the situation were reversed? I also asked them individually how they would want to die if they could have the option of choosing?

All in all, a voluntary decision, not made in haste, thoughtful, careful, meticulous. Decision made as arrangements were made for my passing including burial arrangements, financial and family orders.

Decision also heavily considered with spiritual advisors (chaplain, Rabbi, etc.)

Challenges faced:

MAID in general is highly controversial and its use is RARE – even for patients who do receive it, many do not end up using it. Only a handful of physicians who support using it. Makes it unknown and scary for physicians and patients alike; limited research

Makes acceptance of its use more difficult for family members, too

Prescribing MAID (for some physicians) may feel counter to physician identity as healer & fixer; may spark deep internal/ethical/moral debate for individual physicians as they weigh the option of whether to prescribe

Do they see this as an act of compassion for patients who wish to relieve their suffering?

Do they see this as prescribing a means of suicide?

Anorexia specific – for me, a big issue that caused most ethical debate was whether my case of anorexia nervosa was “reversible.” Many physicians misunderstand SEAN (not even an official DSM diagnosis) and that while anorexia nervosa is a psychiatric illness, it comes with severe medical complications that ultimately are the reason for death. Some of the physicians I worked with could not believe my illness was indeed terminal, but rather felt that there would be something that could be done to reverse the physical damage done to my body that would somehow lengthen my life (even if not for very long – i.e., 1 year).

Yes, perhaps I could stay alive for a few months while in the hospital, but I would have to live in the hospital (MDs might see the benefit of this, but could I? NO! This is where my own reflection around quality of life came in)

My personal belief that this is what makes having such an extreme form of AN so agonizing – mental and emotional suffering is compounded by painful physical complications

Gross misunderstanding about anorexia nervosa in general.

Just over a day before she died, Alyssa wrote to Dr. G, “Thank you with all of my heart for helping to make this possible. I view it as a tremendous act of love.” With family and spiritual support surrounding her, Alyssa became unresponsive in the natural course of her malnutrition. Shortly thereafter, she passed away peacefully. She never actually ingested the MAID medication she had at her disposal.

By presenting these three cases, we have intended to convey some of the emotional, moral, and ethical challenges and dilemmas that patients with SE-AN, their families, and their professional caregivers may face at the end of life. Suffering from unrelenting and irredeemable disorders, these patients made difficult choices, ultimately deciding “enough is enough” [ 18 ]. The anguish endured by these patients and their families resulted in part from lack of professional understanding and consensus regarding terminal care for patients with AN. Neither the fields of palliative and hospice care nor eating disorders have provided definitions or guidance regarding what constitutes a terminal condition in AN or proper ways to address patients and their families grappling with this condition.

Accordingly, we present the following proposed clinical characteristics of those with terminal AN for consideration by both fields (Table 1 ). As illustrated by our cases, no set of criteria will apply perfectly to every patient who identifies with having a terminal case of AN. However, based on prior literature on criteria for clinical terminality [ 15 ], high SMR in those who have previously received inpatient care, are older, and have a history of more severely medically compromised presentations [ 6 , 7 , 8 , 9 , 10 , 11 ], and clinical expertise, the authors propose these clinical characteristics. Some deviation within the second and third characteristics is to be expected and must be individualized to the patient situation. However, the first and fourth must be met in full.

Proposed clinical characteristics of patients with terminal anorexia nervosa

A diagnosis of anorexia nervosa . Anorexia nervosa is the only eating disorder that carries a guaranteed medical cause of death from malnutrition should weight loss continue unabated. As a result, consistent with literature on duration of life during hunger strikes resulting in death [ 16 ], a prognosis of less than 6 months can fairly be established when the patient acknowledges further treatment to be futile and stops engaging in active recovery work. A less than six-month prognosis is congruent with current practice around determination of terminal diagnoses. We fully recognize that patients with SE-AN are likely to have other psychiatric conditions as well.

Age of 30 or older . This criterion accommodates for what is clinically seen as a potential “late maturation phase” in which even those who have been sick for a long time may discover a shift in values and desires that motivates recovery as they enter their late 20 s. Every effort should be made to promote full recovery and continuation of life in those younger than 30. However, the SMR data of multiple recent studies showing the highest death rates in those with a history of inpatient admissions, longer duration of AN, and age over 30 years old [ 6 , 7 , 8 , 9 ], taken alongside what functionally has often been a decade or two of exhaustive, ultimately unsuccessful eating disorder treatment, indicates that the age of around 30 as a minimum for terminal AN is reasonable.

Prior persistent engagement in high-quality, multidisciplinary eating disorder care. Worldwide access to expert eating disorder care varies widely, as does the availability of access to expert inpatient, residential, and full day treatment programs for those with eating disorders. Thus, the definition of care identified here must remain somewhat broad. Before someone can decide they cannot recover, they must have participated in high-quality, expert care to the maximum extent that this is available. This provision should motivate policies that allow for transfers of patients out of designated “networks” that lack expertise, with funding coverage provided at a center of excellence. Ideally, at least some of this treatment will have been undertaken at a sufficiently high level of care to provide extensive structure and support, preferably to the point of full weight restoration at least once in the relatively recent past. Congruent with receipt of such care, qualified health care professionals on the team must support the patient in their decision to stop fighting. We acknowledge that many factors may impact patients’ ability to participate in such care, including lack of access to eating disorders expertise, limitations of the healthcare system, and a personal sense—often based on prior treatment experiences—that admission to certain care settings would cause more harm than good.

Consistent, clear expression by an individual who possesses decision-making capacity that they understand further treatment to be futile, they choose to stop trying to prolong their lives, and they accept that death will be the natural outcome. Careful determination of decisional capacity is required in each case [ 19 ]. An individual who wavers in their conviction or expresses different goals to different people is not yet ready to receive the appellation of terminal AN.

Most eating disorders providers have cared for patients with AN who, despite suffering for decades, continue to show extraordinary determination and resilience. These patients still want help, at least with a harm-avoidance strategy if not with outright full recovery. In these cases, every effort must be made to support the patient’s wishes and provide appropriate resources for recovery. There must be no “giving up” on those who still seek to get better. Indeed, the drive to live and ability to find aspects of life worth fighting for can be seen vividly in the majority of those with AN, even in the face of years or decades of illness and suffering. The psychological imperatives of AN that often lead patients to resist or refuse clinically appropriate care, hazarding medical and psychological risk and deterioration, may seem to conflict with a stated desire to keep trying for recovery. However, in honoring patient autonomy, responsive care must always be offered as long as an individual states that this is their wish.

Patients in their earlier and younger years of AN may say they would rather die than gain weight or nourish themselves properly, a characteristic indicating that AN may present as an ego-syntonic mental illness. Nonetheless, the majority of patients with AN ultimately recover, and such expressions of anguish can be met with compassion and appropriate multidisciplinary care. We would not condone accepting a terminal diagnosis in younger patients. Of note, there are no explicit physiologic markers or measurables (weight, degree of weight loss, presence of or degree of organ failure, vital signs) which delineate someone with terminal AN. Even individuals with extreme medical malnutrition may recover fully if they so choose and have access to expert care. By contrast, if all criteria for terminal AN are met, as in the case of Aaron, individuals should not be obliged to demonstrate extreme medical instability before having the right to choose to stop fighting. Furthermore, while the obsessional ruminations of individuals with AN can be perplexing, clinicians should not regard the presence of body distortions and food fears as proof that these patients are unable to understand personal options and make reasoned health care decisions.

How can we determine that patients with severe anorexia nervosa possess the clinical decision-making capacity necessary to permit them to withdraw from treatment? With respect to decision-making capacity, four traditional criteria are usually applied: understanding, appreciation, ability to reason, and communication of decision [ 20 ]. In Dr. G’s estimation, confirmed in the two cases where formal independent assessment by a psychiatrist was performed, each of the patients met these criteria and was therefore capable of deciding to withdraw from conventional treatment. Alyssa’s clear, incisive writing just days before her death beautifully illustrates the insight and cognitive capacity that many patients with AN possess right up to the end of their life.

Clinical, legal, and ethical commentators in the field concur that withdrawal from treatment may be appropriate when further treatment, whether voluntary of involuntary, will provide only brief improvement, and is unlikely to offer sustained quality of life [ 21 , 22 ]. A formal assessment of decision-making capacity may help ameliorate family member fears that such an important decision is being made in an appropriate and ethical manner, especially when AN fears and distortions can seem so irrational. In addition, a formal bioethics evaluation might be valuable, but consideration of this must be balanced against most bioethicists’ lack of experience with patients who have AN, with the risk that their own innate and misguided reaction that “this patient just has to eat” could undermine a qualified patient’s decisions that are supported by their longstanding care team and family. Even medical ethicists must be wary about how their own cognitive and affective biases might influence their recommendations. [ 23 ]

Family members and carers play an immensely important role in the lives of those with AN. They bear witness to the suffering and challenges experienced by those with AN and are usually directly involved in the recovery process in multiple ways (financial/material support/behavioral support/engagement in the therapeutic work, among others). Many dread the day their child legally becomes an adult and can choose to exclude them from the details of recovery work, such that they become the financial supporters of care they are no longer privy to. The exhaustion, fear, love, and hope experienced by family members cannot be overstated. In any case where a patient meets the criteria for terminal AN, it is always preferable to include family members in the discussions and ideally come to a consensus. There may be dissent within a family about whether their loved one should be allowed to make the decision to stop fighting. These three cases illustrated how each family was meaningfully involved in the clinical discussions in the months before each patient’s death. Each family’s ultimate acceptance (through deep grief) of their son or daughter’s prognosis and choice contributed to a heightened sense of connection and love prior to death.

Acknowledging the considerable controversies surrounding MAID for patients with mental disorders [ 24 , 25 , 26 ], we also submit that patients with terminal AN who are severely physiologically compromised, and whose end-of life suffering results from both psychological and physical pain, should be afforded access to medical aid in dying in locations where such assistance has been legalized—just like other patients with terminal conditions.

AN confers an exceptionally high death rate. The lack of acceptance of terminality in AN and the absence of professionally condoned protocols and standard procedures for supporting patients and families through these phases further complicates end-of-life stages for the adults with AN who cannot keep fighting. These represent a small fraction even of the population of those with SE-AN. Per our proposed clinical characteristics, patients must not only decline further recovery-oriented treatment (which is not uncommon at times for those with AN), but also must explicitly and consistently choose to stop trying to prolong their lives, accepting that death will be the natural outcome. When a patient begins talking about the possibility of not being able to survive, every effort should be made to validate such a serious perspective and to offer an individualized and thoughtful series of harm reduction strategies and treatment options that might make life bearable. However, the process of seeking alternatives to death must not be so exhaustive as to disrespect limits the patient sets; while a family might be desperate for their loved one to try an experimental treatment or “just try going to treatment one more time,” they must ultimately accept the patient’s lack of consent for these.

Our proposed clinical characteristics of patients with terminal AN have no bearing on those who wish to keep fighting despite very long-standing and severe disease, even when their eating disorder behaviors seem incongruent with survival. Very specifically, to move toward a designation of terminal AN, an individual must express consistently that they can no longer live with their disease and will no longer maintain a minimum nutritional intake needed to support life. To be clear, each patient is unique and requires careful individual assessment and consideration as to the best approach going forward. Consistent with calls from others regarding the need for better definition and agreement regarding labeling and staging for SE-AN in general [1,2,3,4 5], the authors hope that these cases and characteristics of those with terminal AN will provide a starting point for identification, care, and further discussion. We would strongly encourage the development of expert consensus criteria and clinical guidelines endorsed by both the fields of palliative and hospice care and eating disorders. These brave, suffering individuals deserve no less.

Availability of data and materials

All narrative data and record of e-mails exchanged with patients and families throughout and after their care with the clinic are available.

Abbreviations

• Anorexia nervosa

Dr. Jennifer Gaudiani

Medical Aid in Dying

Obsessive Compulsive Disorder

• Severe and enduring anorexia nervosa

Broomfield C, Stedal K, Touyz S, Rhodes P. Labeling and defining severe and enduring anorexia nervosa: a systematic review and critical analysis. Int J Eat Disord. 2017;50(6):611–23.

Article   Google Scholar  

Treasure J, Stein D, Maguire S. Has the time come for a staging model to map the course of eating disorders from high risk to severe enduring illness? An examination of the evidence. Early Interv Psychiatry. 2015;9(3):173–84.

Touyz S, Hay P. Severe and enduring anorexia nervosa (SE-AN): in search of a new paradigm. J Eat Disord. 2015;3:26.

Wonderlich SA, Bulik CM, Schmidt U, Steiger H, Hoek HW. Severe and enduring anorexia nervosa: Update and observations about the current clinical reality. Int J Eat Disord. 2020;53(8):1303–12.

Broomfield C, Noetel M, Stedal K, Hay P, Touyz S. Establishing consensus for labeling and defining the later stage of anorexia nervosa: A Delphi study. Int J Eat Disord. 2021. https://doi.org/10.1002/eat.23600 .

Article   PubMed   Google Scholar  

Guinhut M, Godart N, Benadjaoud MA, Melchior JC, Hanachi M. Five-year mortality of severely malnourished patients with chronic anorexia nervosa admitted to a medical unit. Acta Psychiatr Scand. 2021;143(2):130–40. https://doi.org/10.1111/acps.13261 ( PMID: 33247947 ).

van Eeden AE, van Hoeken D, Hoek HW. Incidence, prevalence and mortality of anorexia nervosa and bulimia nervosa. Curr Opin Psychiatry. 2021;34(6):515–24. https://doi.org/10.1097/YCO.0000000000000739 .

Article   PubMed   PubMed Central   Google Scholar  

Nielsen S, Vilmar JW. What can we learn about eating disorder mortality from eating disorder diagnoses at initial assessment? A Danish nationwide register follow-up study using record linkage, encompassing 45 years (1970–2014). Psychiatry Res. 2021;303: 114091. https://doi.org/10.1016/j.psychres.2021.114091 .

Iwajomo T, Bondy SJ, de Oliveira C, Colton P, Trottier K, Kurdyak P. Excess mortality associated with eating disorders: population-based cohort study. Br J Psychiatry. 2021;219(3):487–93. https://doi.org/10.1192/bjp.2020.197 .

Arcelus J, Mitchell AJ, Wales J, Nielsen S. Mortality rates in patients with anorexia nervosa and other eating disorders. A meta-analysis of 36 studies. Arch Gen Psychiatry. 2011;68(7):724–31.

Smith AR, Zuromski KL, Dodd DR. Eating disorders and suicidality: what we know, what we don’t know, and suggestions for future research. Curr Opin Psychol. 2017;12(22):63–7.

Google Scholar  

Lopez A, Yager J, Feinstein RE. Medical futility and psychiatry: palliative care and hospice care as a last resort in the treatment of refractory anorexia nervosa. Int J Eat Disord. 2010;43(4):372–7.

PubMed   Google Scholar  

Westermair AL, Buchman DZ, Levitt S, Trachsel M. Palliative Psychiatry for severe and enduring anorexia nervosa includes but goes beyond harm reduction. Am J Bioeth. 2021;21(7):60–2.

Westmoreland P, Mehler PS. Caring for patients with severe and enduring eating disorders (SEED): certification, harm reduction, palliative care, and the question of futility. J Psychiatr Pract. 2016;22(4):313–20.

Bachar E, Latzer Y, Canetti L, Gur E, Berry EM, Bonne O. Rejection of life in anorexic and bulimic patients. Int J Eat Disord. 2002;31(1):43–8.

Yager J. Managing patients with severe and enduring anorexia nervosa: when is enough, enough? J Nerv Ment Dis. 2020;208(4):277–82.

Hui D, Nooruddin Z, Didwaniya N, Dev R, De La Cruz M, Kim SH, Kwon JH, Hutchins R, Liem C, Bruera E. Concepts and definitions for “actively dying,” “end of life,” “terminally ill,” “terminal care,” and “transition of care”: a systematic review. J Pain Symptom Manag. 2014;47(1):77–89.

Altun G, Akansu B, Altun BU, Azmak D, Yilmaz A. Deaths due to hunger strike: post-mortem findings. Forensic Sci Int. 2004;146(1):35–8. https://doi.org/10.1016/j.forsciint.2004.03.022 .

Yager J, Ganzini L, Nguyen DH, Rapp EK. Working with decisionally capable patients who are determined to end their own lives. J Clin Psychiatry. 2018;79(4):17r11767.

Appelbaum PS. Intuition, self-reflection, and individual choice: considerations for proposed changes to criteria for decisional capacity. Philos Psychiatry Psychol. 2017;24(4):325–8.

Rebecca Dresser (March 1984/April, 1984) Article and Commentary on Anorexia Nervosa: Feeding The Hunger Artists: Legal Issues In Treating Anorexia Nervosa. Wisconsin Law Review, 1984, 297. https://advance-lexis-com.proxy.hsl.ucdenver.edu/api/document?collection=analytical-materials&id=urn:contentItem:3S41-1KG0-00CW-H0MM-00000-00&context=1516831 .

Hébert PC, Weingarten MA. The ethics of forced feeding in anorexia nervosa. CMAJ. 1991;144(2):141–4.

PubMed   PubMed Central   Google Scholar  

Rhodes R, Strain JJ. Affective forecasting and its implications for medical ethics. Camb Q Healthc Ethics. 2008;17(1):54–65.

Freeland A, Godkin D, Dembo J, Chan P, Knoops F, Lachmann M, Morissette L, Smith DH, Stewart DE, Trew M, Wong MR, Charbonneau M. Medical assistance in dying (MAiD) for persons whose sole underlying medical condition is a mental disorder: challenges and considerations. Can J Psychiatry. 2021;31:7067437211043315.

Sheehan K, Gaind KS, Downar J. Medical assistance in dying: special issues for patients with mental illness. Curr Opin Psychiatry. 2017;30(1):26–30.

Rooney W, Schuklenk U, van de Vathorst S. Are concerns about irremediableness, vulnerability, or competence sufficient to justify excluding all psychiatric patients from medical aid in dying? Health Care Anal. 2018;26(4):326–43.

Download references

Acknowledgements

Dr. Gaudiani would like to acknowledge the Gaudiani Clinic’s nurse, Abby Brockman, RN, for her excellent clinical care of these patients and their families.

There was no funding received for this paper.

Author information

Authors and affiliations.

CEDS-S, FAED, Gaudiani Clinic, Denver, CO, USA

Jennifer L. Gaudiani

Department of Psychiatry, University of Colorado School of Medicine, Aurora, CO, USA

Alyssa Bogetz & Joel Yager

You can also search for this author in PubMed   Google Scholar

Contributions

Dr. Gaudiani was the internist for the three patients, drafted the article, and reviewed and approved revisions. Ms. Bogetz (deceased patient) contributed to the text and proposed criteria. Dr. Yager consulted with Alyssa and her family, contributed to the text, and provided extensive editing.

Corresponding author

Correspondence to Jennifer L. Gaudiani .

Ethics declarations

Ethical approval and consent to participate.

Consent for publication of these cases was provided by all of the deceased patients’ families and by Alyssa prior to her death, and all families asked that patients’ first names be used in lieu of initials. This case report did not require further ethical approval.

Consent for publication

All three authors consented to publish this article. Ms. Bogetz (posthumous author) consented prior to her death, both verbally and in writing. Written informed consent for publication of their clinical details was obtained from the parents of the other two patients. A copy of the consent forms is available for review by the Editor of this journal as an additional file.

Competing interests

The authors have no competing interests to declare with regards to any of the material in this paper.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Alyssa Bogetz: 1985–2021, former patient

^Alyssa Bogetz: deceased

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Gaudiani, J.L., Bogetz, A. & Yager, J. Terminal anorexia nervosa: three cases and proposed clinical characteristics. J Eat Disord 10 , 23 (2022). https://doi.org/10.1186/s40337-022-00548-3

Download citation

Received : 19 October 2021

Accepted : 01 February 2022

Published : 15 February 2022

DOI : https://doi.org/10.1186/s40337-022-00548-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Medical aid in dying
• Obsessive compulsive disorder

Journal of Eating Disorders

ISSN: 2050-2974

• General enquiries: [email protected]

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• My Bibliography
• Collections
• Citation manager

Save citation to file

Email citation, add to collections.

• Create a new collection
• Add to an existing collection

Add to My Bibliography

Your saved search, create a file for external citation management software, your rss feed.

• Search in PubMed
• Search in NLM Catalog
• Add to Search

Toward an understanding of risk factors for anorexia nervosa: a case-control study

Affiliation.

• 1 Department of Psychiatry, Columbia University College of Physicians and Surgeons, NY, NY 10032, USA. [email protected]
• PMID: 18070371
• PMCID: PMC2669537
• DOI: 10.1017/S0033291707002310

Background: Prospective, longitudinal studies of risk factors for anorexia nervosa (AN) are lacking and existing cross-sectional studies are generally narrow in focus and lack methodological rigor. Building on two studies that used the Oxford Risk Factor Interview (RFI) to establish time precedence and comprehensively assess potential risk correlates for AN, the present study advances this line of research and represents the first case-control study of risk factors for AN in the USA.

Method: The RFI was used for retrospective assessment of a broad range of risk factors, while establishing time precedence. Using a case-control design, 50 women who met DSM-IV criteria for AN were compared to those with non-eating disorder DSM-IV psychiatric disorders (n=50) and those with no psychiatric disorder (n=50).

Results: Women with psychiatric disorders reported higher rates of negative affectivity, maternal and paternal parenting problems, family discord, parental mood and substance disorder, and physical and sexual abuse than women with no psychiatric disorder. Women with AN specifically reported greater severity and significantly higher rates of negative affectivity, perfectionism and family discord, and higher parental demands than women with other psychiatric disorders. The role of weight and shape concerns was most salient in the year preceding onset of AN.

Conclusions: Convergent data identifying common risk factors as well as those more severe in the development of AN are emerging to inform longitudinal risk factor and prevention studies for this disorder.

PubMed Disclaimer

Similar articles

• Risk factors and antecedent life events in the development of anorexia nervosa: a Portuguese case-control study. Machado BC, Gonçalves SF, Martins C, Hoek HW, Machado PP. Machado BC, et al. Eur Eat Disord Rev. 2014 Jul;22(4):243-51. doi: 10.1002/erv.2286. Epub 2014 Feb 27. Eur Eat Disord Rev. 2014. PMID: 24577737
• Anorexia nervosa versus bulimia nervosa: differences based on retrospective correlates in a case-control study. Machado BC, Gonçalves SF, Martins C, Brandão I, Roma-Torres A, Hoek HW, Machado PP. Machado BC, et al. Eat Weight Disord. 2016 Jun;21(2):185-97. doi: 10.1007/s40519-015-0236-6. Epub 2015 Oct 27. Eat Weight Disord. 2016. PMID: 26506924
• Risk factors for anorexia nervosa and bulimia nervosa in Japan and compared to a U.S. sample. Pike KM, So M, Hilbert A, Maekawa H, Shimanouchi T, Wilfley D, Dohm FA, Fairburn CG, Weissman RS. Pike KM, et al. Int J Eat Disord. 2021 Feb;54(2):155-167. doi: 10.1002/eat.23442. Epub 2020 Dec 23. Int J Eat Disord. 2021. PMID: 33355934
• Recent Advances in Developmental and Risk Factor Research on Eating Disorders. Bakalar JL, Shank LM, Vannucci A, Radin RM, Tanofsky-Kraff M. Bakalar JL, et al. Curr Psychiatry Rep. 2015 Jun;17(6):42. doi: 10.1007/s11920-015-0585-x. Curr Psychiatry Rep. 2015. PMID: 25894358 Review.
• Anorexia Nervosa/Atypical Anorexia Nervosa. Moskowitz L, Weiselberg E. Moskowitz L, et al. Curr Probl Pediatr Adolesc Health Care. 2017 Apr;47(4):70-84. doi: 10.1016/j.cppeds.2017.02.003. Curr Probl Pediatr Adolesc Health Care. 2017. PMID: 28532965 Review.
• Do risk factors differentiate DSM-5 and drive for thinness severity groups for anorexia nervosa? Dang AB, Kiropoulos L, Anderluh M, Collier D, Fernandez-Aranda F, Karwautz A, Treasure J, Wagner G, Krug I. Dang AB, et al. J Eat Disord. 2024 Jan 11;12(1):5. doi: 10.1186/s40337-024-00966-5. J Eat Disord. 2024. PMID: 38212857 Free PMC article.
• Systematic review of prospective studies assessing risk factors to predict anorexia nervosa onset. Charrat JP, Massoubre C, Germain N, Gay A, Galusca B. Charrat JP, et al. J Eat Disord. 2023 Sep 20;11(1):163. doi: 10.1186/s40337-023-00882-0. J Eat Disord. 2023. PMID: 37730675 Free PMC article. Review.
• Anorexia Nervosa in Adolescence: Parental Narratives Explore Causes and Responsibilities. Carpinelli L, Watzlawik M. Carpinelli L, et al. Int J Environ Res Public Health. 2023 Feb 24;20(5):4075. doi: 10.3390/ijerph20054075. Int J Environ Res Public Health. 2023. PMID: 36901086 Free PMC article.
• The Relationship of Parental Personality Disorders with Offspring Eating Disorders at Childhood and Adolescence Age. Pourdehghan P, Mohammadi MR, Mostafavi SA, Khaleghi A, Ahmadi N. Pourdehghan P, et al. Child Psychiatry Hum Dev. 2024 Apr;55(2):361-371. doi: 10.1007/s10578-022-01407-7. Epub 2022 Aug 14. Child Psychiatry Hum Dev. 2024. PMID: 35964272
• Can Cognitive Flexibility and Clinical Perfectionism Be Used to Identify People with Anorexia Nervosa? Miles S, Nedeljkovic M, Phillipou A. Miles S, et al. J Clin Med. 2022 Mar 31;11(7):1954. doi: 10.3390/jcm11071954. J Clin Med. 2022. PMID: 35407563 Free PMC article.
• Anderluh M, Tchanturia K, Rabe-Hesketh S, Treasure J. Childhood obsessive-compulsive personality traits in adult women with eating disorders: defining a broader eating disorder phenotype. American Journal of Psychiatry. 2003;160:242–247. - PubMed
• Bastiani AM, Rao R, Weltzin T, Kaye WH. Perfectionism in anorexia nervosa. International Journal of Eating Disorders. 1995;17:147–152. - PubMed
• Bruch H. The Golden Cage: The Enigma of Anorexia Nervosa. Harvard University Press; Cambridge: 2001.
• Bulik C, Tozzi F, Anderson C, Mazzeo S, Aggen S, Sullivan P. The relation between eating disorders and components of perfectionism. American Journal of Psychiatry. 2003;160:366–368. - PubMed
• Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd edn Erlbaum; Hillsdale, NJ: 1988.

Publication types

• Search in MeSH

Related information

Grants and funding.

• MH52348/MH/NIMH NIH HHS/United States
• R01 MH052348/MH/NIMH NIH HHS/United States
• K24 MH070446/MH/NIMH NIH HHS/United States
• 1K24MH070446/MH/NIMH NIH HHS/United States
• 046386/WT_/Wellcome Trust/United Kingdom
• 079113/WT_/Wellcome Trust/United Kingdom
• MH57533/MH/NIMH NIH HHS/United States

LinkOut - more resources

Full text sources.

• Cambridge University Press
• Europe PubMed Central
• Ovid Technologies, Inc.
• PubMed Central

Other Literature Sources

Miscellaneous.

• NCI CPTAC Assay Portal
• Citation Manager

NCBI Literature Resources

MeSH PMC Bookshelf Disclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Cambridge Open

Anorexia nervosa: 30-year outcome

Sandra rydberg dobrescu.

1 Student, Gillberg Neuropsychiatry Centre, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Sweden

Lisa Dinkler

Carina gillberg.

2 Associate Professor, Gillberg Neuropsychiatry Centre, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Sweden

Maria Råstam

3 Professor of Child and Adolescent Psychiatry, Department of Clinical Sciences Lund, Lund University; and Visiting Professor, Gillberg Neuropsychiatry Centre, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Sweden

Christopher Gillberg

4 Professor of Child and Adolescent Psychiatry, Gillberg Neuropsychiatry Centre, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Sweden

Elisabet Wentz

5 Professor of Psychiatry, Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, Sahlgrenska Academy, University of Gothenburg, Sweden

Associated Data

For supplementary material accompanying this paper visit https://doi.org/10.1192/bjp.2019.113.

Little is known about the long-term outcome of anorexia nervosa.

To study the 30-year outcome of adolescent-onset anorexia nervosa.

All 4291 individuals born in 1970 and attending eighth grade in 1985 in Gothenburg, Sweden were screened for anorexia nervosa. A total of 24 individuals (age cohort for anorexia nervosa) were pooled with 27 individuals with anorexia nervosa (identified through community screening) who were born in 1969 and 1971–1974. The 51 individuals with anorexia nervosa and 51 school- and gender-matched controls were followed prospectively and examined at mean ages of 16, 21, 24, 32 and 44. Psychiatric disorders, health-related quality of life and general outcome were assessed.

At the 30-year follow-up 96% of participants agreed to participate. There was no mortality. Of the participants, 19% had an eating disorder diagnosis (6% anorexia nervosa, 2% binge-eating disorder, 11% other specified feeding or eating disorder); 38% had other psychiatric diagnoses; and 64% had full eating disorder symptom recovery, i.e. free of all eating disorder criteria for 6 consecutive months. During the elapsed 30 years, participants had an eating disorder for 10 years, on average, and 23% did not receive psychiatric treatment. Good outcome was predicted by later age at onset among individuals with adolescent-onset anorexia nervosa and premorbid perfectionism.

Conclusions

This long-term follow-up study reflects the course of adolescent-onset anorexia nervosa and has shown a favourable outcome regarding mortality and full symptom recovery. However, one in five had a chronic eating disorder.

Follow-up studies of anorexia nervosa have been conducted since the second half of the 20th century and were thoroughly reviewed in 2009 by Steinhausen. 1 According to the review, approximately half of all individuals with anorexia nervosa were classified as fully recovered, one in three had improved and one in five had a chronic course of the disorder. The crude mortality rate was 5%. 1 High mortality rates were also found in a meta-analysis reporting a standardised mortality ratio of six for anorexia nervosa. 2 According to Steinhausen, the outcome of adolescent-onset anorexia nervosa was more favourable in terms of mortality and chronicity than the outcome of anorexia nervosa with variable onset, including patients with adult onset. 1 All 119 patient series in the Steinhausen review 1 except two – the FinnTwin study 3 and the present study, i.e. the ‘Gothenburg anorexia nervosa study’, 4 – 7 – were based on clinical data.

Three anorexia nervosa outcome studies (one from Germany, 8 one from the USA 9 and one from Sweden 10 ) report exceptionally long observational periods of more than 20 years. All three studies included patients only. Recovery from anorexia nervosa was observed in 51% of patients after 21 years in the German study, 8 in 63% of patients after 22 years in the American study 9 and in 76% of patients after 33 years in the Swedish study. 10 Two of the studies reported crude mortality rate, corresponding to 16 and 18%, respectively. 8 , 10 The three outcome studies either performed personal interviews 8 , 10 or telephone interviews 9 at follow-up. Two of the studies had a prospective design. 8 , 9

Since 1985, we have carried out a prospective, longitudinal, case–control study of individuals with adolescent-onset anorexia nervosa. The individuals have been examined on four previous occasions. The aims of the present study were as follows:

• to prospectively examine the very long-term outcomes of adolescent-onset anorexia nervosa, including full recovery from eating disorder symptoms, psychiatric morbidity, mortality, global functioning and health-related quality of life (HRQoL).
• to identify predictors of outcome to determine risk factors for developing anorexia nervosa.

For our study group we hypothesised that the outcomes of Global Assessment of Functioning (GAF), HRQoL, eating disorder outcome and psychiatric morbidity would be significantly worse than for the matched comparison group. We hypothesised that the outcomes within our study group would be better than the outcomes of the clinical, long-term studies due to this sample being partly community and population based and only including individuals with adolescent-onset anorexia nervosa.

Study design and participants

The original study (study 1), the anorexia nervosa group.

In 1985, the so-called Gothenburg anorexia nervosa study was initiated by M.R. and C.G. All 4291 individuals born in 1970 and attending eighth grade in Gothenburg underwent a physical examination and completed an eating disorder symptom questionnaire. All 4291 growth charts and questionnaires were scrutinised. All individuals suspected to have anorexia nervosa were examined in person. All children in special schools were assessed at the same time. Regarding all boarding schools, M.R. established contact with the school nurses and interviewed them regarding any child born in 1970 who could have anorexia nervosa. There were no cases of home education. In the 1970 birth cohort, 23 girls and 2 boys met criteria for anorexia nervosa before age 18. Only 1 girl declined further examination, leaving 24 individuals (22 girls, 2 boys) in the ‘population-based group’ (for further details see 7 ). Three individuals from the population-based group fulfilled almost all the criteria for anorexia nervosa according to the DSM-III-R (1987); these participants were considered as partial cases. These three individuals met all criteria for anorexia nervosa shortly after study 1 (see below). At the time of the epidemiological study, individuals with anorexia nervosa, born in 1969 and 1971–1974, were also reported to the researchers by the school health services after community screening. In all, 27 adolescents (26 girls, 1 boy) formed the ‘population-screening group’, which was not a clinical group. A third of the population-screening group had not been in touch with any paediatric, child and adolescent psychiatric or adult psychiatric service.

Due to the similar group structure, the two groups were pooled together to form the ‘anorexia nervosa group’, consisting of 51 participants with anorexia nervosa (48 girls, 3 boys). 7 All 51 individuals had been assigned a diagnosis of anorexia nervosa by a physician specialised in psychiatry (M.R.) and all met the DSM-III-R 11 and the DSM-IV (1994) 12 criteria for anorexia nervosa. Therefore, all participants were found to be true cases of the disorder. The mean age at onset of anorexia nervosa was 14.3 years (range 10.0–17.2 years).

Comparison group

At the time of the anorexia nervosa screening, a comparison group was recruited. For each of the 51 participants with anorexia nervosa, the school health nurses selected a same-gender classmate who was closest in age to the index child. These comparison participants had no history of an eating disorder. This group also consisted of 51 individuals (48 girls, 3 boys).

At the time of the original study (study 1), all 102 individuals (51 anorexia nervosa, 51 comparison) underwent a physical and psychiatric examination. The mothers were interviewed in a semi-structured way and they completed questionnaires about the child's developmental history, psychiatric and personality traits, temperament, interests and peer relationships. The assessment of childhood perfectionism was made as a clinical judgement based on several-hours-long interviews with the parents (e.g. including a semi-structured validated interview). 13 Information from a parental questionnaire, the child healthcare centre and the patient registers were also used.

M.R. compiled case notes for all individuals. All information that could raise suspicion that the child might have an eating disorder had been excluded in these records. C.G., who was blinded to group status, assigned psychiatric diagnoses (including autism spectrum disorder [ASD]) and personality traits (including perfectionism) based on the data from the case notes. 7 In the anorexia nervosa group, 12 participants had developed bulimia nervosa after the onset of anorexia nervosa and a diagnosis of bulimia nervosa was met either before or at study 1. This information was based on growth charts, school nurse assessments and questionnaires (and reported at study 1). 7

Study 2, study 3 and study 4

The anorexia nervosa and comparison groups were followed up 6 years after the onset of anorexia nervosa (study 2; mean age 21 years), 10 years after onset (study 3; mean age 24 years) and after 18 years (study 4; mean age 32 years). All 102 individuals agreed to participate in all follow-up studies.

Study 5 (the present study)

All 102 participants were traced. Four individuals in the anorexia nervosa group (two women, two men) declined participation (one man lived abroad and worked ‘12–15 h a day’, one woman was too disabled due to an anxiety disorder and one man and one woman were not comfortable taking part in the research project), leaving 47 participants (46 women, 1 man; 36 face-to-face interviews, 11 online video conferences/telephone interviews) in the anorexia nervosa group. All 51 individuals (48 women, 3 men; 42 face-to-face interviews, 9 online video conferences/telephone interviews) in the comparison group agreed to participate. In all, 98 out of 102 individuals participated in the study (drop-out rate 4%).

This study was approved by the Regional Ethical Review Board at the University of Gothenburg (398–14) and followed the World Medical Association's Declaration of Helsinki. The individuals participated voluntarily. Written informed consent was obtained from all participants.

The Mini-International Neuropsychiatric Interview (MINI 6.0) 14 was used. Two of the individuals in the anorexia nervosa group only agreed to participate in a short interview and therefore the MINI was conducted to 45 out of the 47 participants in the anorexia nervosa group. The eating disorder module of the MINI was considered too brief, and therefore the eating disorder domain of the Structured Clinical Interview for DSM-IV (SCID-I) 15 and a DSM-5 (2013) checklist for feeding and eating disorders were added. In the present study the DSM-5 criteria for feeding and eating disorders have been applied. 16

The GAF was used to assess general outcome. The Morgan–Russell scales were used to calculate a ‘Morgan–Russell averaged scale score’, which is a composite score summarising body and weight concern, dieting, body weight, menstrual status, mental state, attitudes toward sexual relationships and menstruations, social relationships with family and friends, and employment during the past 6 months. The Morgan–Russell scales are well established and the best-validated outcome instruments in anorexia nervosa research. 17

The 36-item Short Form Health Survey (SF-36), 18 a generic HRQoL instrument, is the most frequently applied HRQoL measurement in eating disorders. 19 It consists of four physical and four mental subscales. The instrument includes two composite scores: the Physical Composite Score (PCS) and the Mental Composite Score (MCS).

Full recovery from eating disorder symptoms

In accordance with our definition of full recovery from eating disorder symptoms in study 3 and study 4, 5 , 6 informed by Strober et al , 20 a fully recovered individual ‘refers to patients who have been free of all criterion symptoms of anorexia nervosa or bulimia nervosa for not less than 8 consecutive weeks’. It requires ‘the sustained absence of weight deviation, compensatory behaviours, and deviant attitudes regarding weight and shape, including weight phobia’. In our sample, we also required being free from all criteria of binge-eating disorder and that the individuals had been free from the above symptoms for a minimum of 6 months. In the present study information was obtained from the MINI, SCID-I, the DSM-5 checklist for feeding and eating disorders and the Morgan–Russell scales.

Statistical analyses

Mainly non-parametric tests were used, due to the scale scores not being normally distributed. The chi-squared test or Fisher's exact test were used for dichotomous variables and the Mann–Whitney test for continuous variables. To assess change over time, McNemar's test and the Wilcoxon signed-rank test were used for dichotomous and continuous variables, respectively. A P -value of <0.05 was regarded as statistically significant. In cases of multiple comparisons, the upper limit of false significance was calculated as follows: (number of tests − number of significant tests on significance level 0.05)*0.05/(1–0.05).

Predictive factors

Predictive factors were chosen in accordance with the 18-year follow-up study (study 4) 6 and focused mainly on variables measured at study 1 and on premorbid/retrospective childhood data. The following predictors were investigated for their individual relationships with continuous outcome variables (GAF, Morgan–Russell averaged scale score, PCS, MCS) and the dichotomous outcome variable of full recovery from eating disorder symptoms: age at onset of adolescent-onset anorexia nervosa and premorbid body mass index (BMI); lowest BMI ever assessed at study 2 (continuous); perinatal factors and social class (ordinal); early gastrointestinal problems; primary amenorrhoea; premorbid problems including problems with friends, major problems in the family, major life events, parental divorce, death of family member; affective disorder; overweight; obsessive–compulsive disorder; obsessive–compulsive personality traits; perfectionism; and ASD before study 1 and/or at study 2 (all dichotomous). Spearman correlations were used to examine the association of two continuous/ordinal variables, Mann–Whitney test for the association of continuous/ordinal with dichotomous variables and logistic regression for the univariate associations of predictors with full recovery of eating disorder symptoms. Variables with P  < 0.10 in the univariate tests were included as possible predictors in the multiple stepwise regression analyses (‘backward’ procedure in SPSS).

There was no mortality. The mean ages at study 5 and the length of the follow-up periods are presented in Supplementary Table S1 available at https://doi.org/10.1192/bjp.2019.113 . Anthropometric data are shown in Table 1 .

Comparisons between the anorexia nervosa and the comparison group regarding outcome variables and changes between baseline and study 5, and between study 4 and study 5

‘Baseline’ corresponds to study 1 (the original study), when the anorexia nervosa screening was performed. Average minimum BMI at the time was 14.9 kg/m 2 (s.d. 2.6) in the anorexia nervosa group. Out of 51 individuals in the anorexia nervosa group, 12 no longer fulfilled an anorexia nervosa diagnosis at the time of study 1. Due to multiple comparisons the upper limit of false significance was calculated to be 1.2. Study 4, 18-year follow-up; study 5, 30-year follow-up/the present study; n/a, not applicable; GAF, Global Assessment of Functioning; BMI, body mass index (kg/m 2 ).

General outcome

The outcome variables of the GAF and Morgan–Russell averaged scale score were significantly lower in the anorexia nervosa group ( Table 1 ).

Changes in outcome between study 4 and study 5

Weight and BMI had increased significantly in the anorexia nervosa and comparison group between study 4 and study 5 (Supplementary Table S1). The GAF and Morgan–Russell averaged scale score had not changed between the last two follow-ups in either group.

Eating disorders at study 5

Nine individuals (19%) in the anorexia nervosa group had a current eating disorder, including three people with anorexia nervosa (two classified as being in partial remission. If DSM-IV criteria had been applied, one of the three people with anorexia nervosa would have been classified as eating disorder not otherwise specified due to regular menstruations) ( Table 2 ). The mean duration of the first episode of anorexia nervosa (calculated from onset of anorexia nervosa) was 3.6 years (s.d. 3.0). The mean duration of all aggregated episodes of anorexia nervosa was 4.9 years (s.d. 5.1). The mean duration of all aggregated episodes of eating disorders (including anorexia nervosa) was 10.2 years (s.d. 8.1).

The prevalence of psychiatric diagnoses at study 5 and the prevalence of psychiatric diagnoses between study 4 and study 5

For feeding and eating disorders the DSM-5 criteria have been applied; criteria for other psychiatric disorders were based on the DSM-IV (Mini-International Neuropsychiatric Interview [MINI] 6.0). Study 4, 18-year follow-up; study 5, 30-year follow-up/the present study; OSFED, other specified feeding or eating disorder; OCD, obsessive–compulsive disorder.

* P  < 0.05, ** P  < 0.01, *** P  < 0.0001; anorexia nervosa v. comparison group.

In the anorexia nervosa group, 64% ( n  = 30) were considered fully recovered. There was no significant difference in mean BMI between recovered and non-recovered participants (recovered: 22.7 kg/m 2 , s.d. 3.1; non-recovered: 24.5 kg/m 2 , s.d. 7.2; P  = 0.24). The three individuals with current anorexia nervosa, including the two in partial remission, had a mean BMI of 19.7 kg/m 2 (s.d. 3.1, range 16.3–22.4). Those with full recovery of eating disorder symptoms had significantly better outcome according to the GAF and Morgan–Russell averaged scale score than those without full eating disorder symptom recovery.

Changes in eating disorder diagnoses between study 4 and study 5 and over 30 years

Between the two most-recent follow-up studies, 17% ( n  = 8) of participants fulfilled the criteria for anorexia nervosa at some point and 32% ( n  = 15) had experienced any type of eating disorder (including anorexia nervosa) ( Table 2 ; Supplementary Figure S1). Figure 1 shows the trajectories of the individuals' eating disorders over 30 years. All of those who had anorexia nervosa at study 4 were now free from any eating disorder. Figure 2 shows BMI over 30 years.

The trajectories of eating disorders over 30 years. The eating disorder diagnoses at each of the five assessments for all 51 individuals in the anorexia nervosa group is shown. The numbers before the abbreviations in the boxes correspond to the number of individuals with the condition. The column to the left shows the number of the study and, within brackets, the mean age of the anorexia nervosa group at the time of the study. From study 1 to study 4 the eating disorder diagnoses were assigned according to the DSM-IV; the DSM-5 criteria were applied at study 5. AN, anorexia nervosa; BE, binge-eating disorder; BN, bulimia nervosa; DO, dropped out; ED, eating disorder not otherwise specified (other specified feeding or eating disorder according to the DSM-5); NO, no eating disorder.

Body mass index (BMI) in the anorexia nervosa and comparison group in the original study and across the four follow-up studies. Average BMI in each group at each of the five assessments is displayed. The number below each study indicates the mean age of all participants at that assessment. Error bars indicate 95% CI. AN, anorexia nervosa; AN study 1, the original study; AN study 2, 6-year follow-up; AN study 3, 10-year follow-up; AN study 4, 18-year follow-up; AN study 5, 30-year follow-up (the present study); COMP, comparison.

Other psychiatric disorders at study 5

Psychiatric morbidity was significantly over-represented in the anorexia nervosa group ( Table 2 ). DSM-5 anxiety disorders were the most common psychiatric disorders in both groups, but they were significantly more common in the anorexia nervosa group ( Table 2 ).

Changes in other psychiatric disorders between study 4 and study 5 and over 30 years

Between the two most-recent follow-up studies, psychiatric diagnoses were significantly more common in the anorexia nervosa group than in the comparison group ( Table 2 ). The percentages of psychiatric disorders in the original study and across the four follow-up studies are available in Figure S2.

Treatment received

In the anorexia nervosa group, 23% had never received treatment for an eating disorder. Only two individuals were in current treatment for an eating disorder at study 5, including one person with anorexia nervosa who received compulsory treatment. According to the mean Morgan–Russell averaged scale score and the mean GAF there was no difference between those who had ever received treatment for an eating disorder and those who had not received treatment.

A total of 37 people from the anorexia nervosa group and 47 from the comparison group completed the SF-36 (Table S2). A drop-out analysis showed that people in the anorexia nervosa group who completed the survey had a significantly lower mean GAF score (58.3) than those who did not (69.8; P  = 0.043), and psychiatric comorbidity was more common in those who completed the survey than in those that did not ( P  = 0.015). The MCS was significantly lower in the anorexia nervosa group (Supplementary Table S2). Neither the mental subscale scores nor the MCS were significantly lower among those in the anorexia nervosa group with a current eating disorder. Individuals in the anorexia nervosa group with a current psychiatric morbidity scored significantly lower in all subscale and composite scores than individuals without a psychiatric disorder.

Correlations between previously diagnosed ASD and other variables

Our four previous studies have shown that 12% ( n  = 6) of the anorexia nervosa group had ASD at all four examinations (ASDx4). 6 At study 5, the mean GAF and Morgan–Russell averaged scale score were significantly lower in the ASDx4 group compared with the remainder of the anorexia nervosa group. No individuals in the ASDx4 group had a current eating disorder at study 5.

Predictive factors for good outcome

Stepwise linear regression analysis revealed that higher age at onset of adolescent-onset anorexia nervosa and perfectionism before onset of anorexia nervosa were individual predictors for better outcome on GAF, Morgan–Russell averaged scale score and MCS. Additionally, early gastrointestinal problems were an individual predictor for better outcome on the Morgan–Russell averaged scale score. In the stepwise logistic regression for full recovery from eating disorder symptoms, higher age at onset of adolescent-onset anorexia nervosa was a significant predictor (Tables S3 and S4).

This unique, controlled, 30-year follow-up study of adolescent-onset anorexia nervosa has shown that the majority of people with this disorder do well in the long-term perspective. There was no mortality, almost two-thirds reported full recovery from eating disorder symptoms and physical aspects of quality of life were similar across the anorexia nervosa and comparison groups. However, eating disorders were still present in a minority, affecting one in five.

Due to the sample being partly population based and including only cases of adolescent-onset anorexia nervosa, we had hypothesised that the outcome of our anorexia nervosa group would be better than other very long-term outcome studies. With no deaths in our sample, we had better mortality results than clinical outcome studies. Our findings were in line with the community-based FinnTwin study, which reported no mortality after 10 years. 21 The proportion of full recovery from eating disorder symptoms (64% in our anorexia nervosa group) was not better than that seen in clinical studies with follow-up periods of more than 20 years. 8 – 10

One other study has an observational period of more than 30 years. 10 The sample had been admitted to hospital during the period 1931–1960. Given that no evidence-based treatment for anorexia nervosa had been developed during that period, it is surprising that 76% of the sample had recovered after 33 years. However, the mortality rate of 18% may be an indication of the lack of knowledge about eating disorders in health services at the time. Only 6% of participants had an eating disorder after 33 years. 10 In our 18-year follow-up, we found 12% with an eating disorder. 6 We expected to find, if anything, lower rates at the 30-year follow-up, but instead we saw a small increase. We were surprised to find that a third of the anorexia nervosa group fulfilled an eating disorder diagnosis during the past 12 years. However, this is in line with a recent 20-year follow-up study of in-patients with anorexia nervosa that showed a remission rate of only 39%. In that study, remission was defined as not fulfilling any eating disorder diagnosis (including ‘eating disorder syndrome below the threshold of DSM-IV’) for the past 3 months. 22

The prospective design enabled us to study the individual trajectories of anorexia nervosa and crossover from one eating disorder to another. Between the 18- and 30-year follow-up, one in five had an eating disorder relapse. Diagnostic crossover involving anorexia nervosa, binge-eating disorder and other specified feeding or eating disorder was also common during this follow-up period. At the 10-year follow-up, study 3, 5 half of the anorexia nervosa group had met criteria for bulimia nervosa at some point, which is in line with other long-term follow-up studies. 20 However, from study 4 and onwards bulimia nervosa was no longer observed among the individuals in the anorexia nervosa group.

In this study, one in four people had never received treatment for an eating disorder. Nonetheless, treatment did not affect the outcome 30 years after the onset of anorexia nervosa. However, since the individuals were not randomly allocated to receive or not receive treatment, these outcome results must be interpreted with caution. Our finding is in accordance with a 5-year follow-up study of eating disorders where treatment did not affect the outcome of any eating disorder. 23 Our results may reflect that individuals with anorexia nervosa are reluctant to undergo treatment and there is meagre scientific evidence for anorexia nervosa treatment per se , with least evidence for adult patients. 19 The resistance to recovery posits anorexia nervosa among the most difficult psychiatric disorders to treat.

Regarding HRQoL, the PCS reflected a good outcome in the anorexia nervosa group, but the MCS did not. The lower MCS results could partly be explained by a significant over-representation of psychiatric morbidity and lower mean GAF among those who completed the SF-36 compared with those who did not. In our sample individuals with a current eating disorder did not express worse HRQoL compared with those who had no current eating disorder, whereas Eddy's group 9 found the physical and psychological aspects of HRQoL to be significantly poorer among people with unrecovered anorexia nervosa at the 22-year follow-up. In a community-based study using the SF-36, individuals with a self-reported history of anorexia nervosa had a poorer HRQoL than those who did not have a history of anorexia nervosa. 24

Later age at onset among individuals with adolescent-onset anorexia nervosa predicted good outcome, i.e. adolescent onset is better than childhood onset according to the GAF and the Morgan–Russell averaged scale score. This is consistent with outcome studies of childhood-onset anorexia nervosa, 25 where the illness ‘often takes a chronic and disabling course with high morbidity rates’. 25 The individuals in our study with the lowest ages of onset were in the same age range as the oldest children in the outcome studies of childhood-onset anorexia nervosa (which had an inclusion criterion of onset of anorexia nervosa before age 14). 25 Premorbid perfectionism was a favourable prognostic factor even though premorbid perfectionism has often been reported as a risk factor for developing anorexia nervosa. 26 Clinical perfectionism may maintain the eating disorder psychopathology 27 and therefore an eating disorder treatment designed to produce enduring change, enhanced cognitive–behavioural therapy, has included a module for reducing clinical perfectionism. In contrast, perfectionism has been shown to persist in individuals with good outcome and anorexia nervosa recovery. 28 Could it be that the perfectionism that drove the illness was diverted to driving recovery, i.e. perfectionism can both help and hinder in attaining a goal? Early severe gastrointestinal problems also predicted a good outcome in the present study. In our original study, M.R. found that half the individuals in the anorexia nervosa group had an early history of severe gastrointestinal problems. 7 These symptoms may reflect an immature gastrointestinal tract that caused a great deal of concern in childhood and during adolescence and therefore contributed to the development of anorexia nervosa, but was subsequently not a risk factor for perpetuating anorexia nervosa.

Strengths and limitations

To our knowledge, this is the only study that has followed people with adolescent-onset anorexia nervosa and matched comparison participants prospectively for 30 years. The sample is community based and half of the participants with anorexia nervosa constitute a total birth cohort. However, some factors may have influenced the course of the illness. Even if all individuals were initially identified in the community, half the group included a greater proportion of individuals who had been in contact with treatment facilities. The majority of individuals received treatment at some point, as opposed to the FinnTwin study where only half of the participants had been detected in the healthcare system. 3 Even though some individuals in our anorexia nervosa group did not seek treatment, one must bear in mind that identifying and following the individuals prospectively for 30 years could be considered an intervention.

The comparison group was matched for age, gender and school, and has been followed since the original study. Only two other research groups have included a comparison group in their follow-up studies of anorexia nervosa. 29 , 30 The drop-out rate after 30 years was extremely low; 96% agreed to examination for the fifth time in this project. The prospective design made it possible to study each individual's eating disorder trajectory in detail over a period of 30 years.

One of the weaknesses of this study pertains to the sample size: 47 individuals with adolescent-onset anorexia nervosa is a relatively small number compared with other very long-term studies where 84–121 8 – 10 patients were followed up. However, the present study is the only one that can claim representativeness, with its population-/community-based design.

Individuals who decline participation in a follow-up study may represent those with the worst outcome. 9 Our drop-out rate was 4%, corresponding to four individuals of the anorexia nervosa group. We had at least some follow-up information on all non-participants; we had been in touch with all of them and at least half of them worked full time. Based on these data we do not believe that our small group of non-participants represents those with the worst outcome.

Our definition of recovery (full recovery from eating disorder symptoms for a minimum of 6 consecutive months) could be questioned and it may have been better to use a longer time period. Some researchers have argued for a 1-year period without eating disorder behaviours, due to the risk of relapse being greatest within 1 year post-treatment or post-recovery. 20 , 31 Our definition of recovery includes physical (BMI), behavioural (absence of binge eating, compensatory behaviours and restrictive eating) and psychological (body image concerns and fear of weight gain) aspects. Khalsa et al 31 have reported that an instrument measuring psychological symptoms of eating disorders, e.g. the Eating Disorder Examination Questionnaire (EDE-Q), 32 should be included in the assessment, and that individuals who exhibit full eating disorder recovery should score in accordance with community norms. A weakness of our study is that we did not use the EDE-Q as an outcome measure and therefore the more detailed aspects of eating disorder symptoms were not assessed. Each of the four other anorexia nervosa studies with a follow-up period of 20 years or more 8 – 10 , 22 have used their own definition of recovery, with one research group arguing that their more strict definition resulted in a recovery rate of only 39%. 22 According to a recent review of the definition of recovery of anorexia nervosa, Khalsa et al 31 conclude that ‘to date there are no consensus guidelines available’ for research purposes.

Another limitation is that referral to eating disorder specialist treatment by the researchers during the 30-year follow-up period may have affected the outcomes of the study. At study 2, only a total of 37% of the population-based group had received treatment. 33 At study 3, study 4 and study 5 the individuals with current anorexia nervosa did not agree to be referred to an eating disorder service.

Clinical implications

Adolescent-onset anorexia nervosa carries a good ‘lifetime’ prognosis in terms of mortality and anorexia nervosa chronicity. However, high prevalence of anorexia nervosa relapses between the 18-year and the 30-year follow-ups indicates that late relapses occur, even though some individuals had been free from an eating disorder for one or two decades. As clinicians, we must be aware that a substantial minority of patients will continue to need psychiatric expertise for their eating disorder or other psychiatric disorders for many years.

Higher age at onset of adolescent-onset anorexia nervosa predicted better general outcome. This finding implies that – among school health nurses, school doctors, child psychiatrists and paediatricians – more effort needs to be made to detect individuals with a very early onset of anorexia nervosa, i.e. in childhood and early adolescence. With regard to the good outcomes predicted by premorbid perfectionism in this study, premorbid and clinical perfectionism in anorexia nervosa has traditionally been regarded as challenging by clinicians in terms of treatment outcome and prognosis. Clinically we may have to reconsider this dogma, at least in terms of the premorbid traits. However, premorbid perfectionism has to our knowledge not been reported as a good prognostic factor previously and therefore our finding needs to be replicated before taking further clinical actions.

Acknowledgements

The authors thank all participants. We thank Ingrid Vinsa for collecting all the anthropometric data.

C.G. received grant support from the Swedish Research Council (521-2012-1754), AnnMari and Per Ahlqvist Foundation and from government grants under the ALF agreement. E.W. received support from Jane and Dan Olsson Foundations (2015 and 2016-55), Wilhelm and Martina Lundgren Foundation (vet2-73/2014 and 2017-1555) and Petter Silfverskiöld's Memorial Foundation (2016-007). L.D. was supported by Queen Silvia's Jubilee Fund (2016) and the Samariten Foundation (2016-0150). All authors received research support from the Birgit and Sten A. Olsson Foundation for research on mental disabilities.

Supplementary material