You are on page 1of 5


Psychoneuroendocrinology (2006) 31, 1149–1153

Psychoneuroendocrinology of anorexia nervosa
P. Södersten!, C. Bergh, M. Zandian

Section of Applied Neuroendocrinology, Mandometer Clinic, Karolinska Institutet, AB Mando, Novum,
S-141 57 Huddinge, Sweden

Received 2 September 2006; received in revised form 15 September 2006; accepted 15 September 2006

KEYWORDS Summary It is suggested that the symptoms of anorexia nervosa are physiological
Anorexia; responses to starvation. There is no evidence of a neural or non-neural dysfunction
Starvation; that predisposes women for anorexia nervosa and the endocrine and psychological
Leptin; consequences of starvation are reversed once patients have re-learnt how to eat and
Treatment regained a normal body weight. Because variability in the supply of food may be a
common evolutionary condition, it is more likely that body weight is variable than
constant in normal circumstances. The role of the neuroendocrine system in times of
feast and famine is to allow the individual to adopt behavioral strategies as needed
rather than maintaining body weight homeostasis. Treatment of anorexic patients
should aim at reducing their high level of physical activity in order to facilitate eating.
& 2006 Elsevier Ltd. All rights reserved.

1. Introduction p. 123), the evidence supporting the serotonin
hypothesis of anorexia nervosa is weak and, prob-
In many accounts, anorexia nervosa is considered a ably for this reason, outcome of treatments based
multifactorial illness of unknown etiology (reviewed on this hypothesis remains poor. Recently, Walsh et
by Södersten et al., 2006). This leaves the clinician al. (2006) reported that treatment with an SSRI does
without a clue as to the cause of the condition and not prevent relapse in anorexic patients in remis-
with the difficult task of selecting the symptoms to sion. Most of the patients relapsed or dropped out of
be treated. Often symptoms thought to be related the study. Unfortunately, most other treatment
to an alteration in serotonergic neurotransmission methods are at best marginally effective (Bergh
are targeted by using selective serotonin reuptake et al., 2006). Clearly, anorexia nervosa needs to be
inhibitors (SSRI). This is a reasonable starting point examined within a new framework.
because serotonin has a role in both mood and
eating behavior. However, as pointed out in the
interesting paper by Mondelli (2006, this issue, 2. Starvation and anorexia nervosa

!Corresponding author. Tel.: +468 55640602; Pioneering clinicians found no sign of organic
fax: +468 55640610. damage in patients with anorexia nervosa
E-mail address: (P. Södersten). (Gull, 1874) and more recent studies have provided

0306-4530/$ - see front matter & 2006 Elsevier Ltd. All rights reserved.

leptin might are mainly a consequence of nutritional factors and actually stimulate food intake. 1975). For example.. 2005). it was recently pointed out that the relapse (see Haas et al. Consider parenteral nutri- published recently (e. normally mediates between body tually will ‘‘give us a window on to the detailed fat and the hypothalamic–pituitary–gonadal axis. This observation ogy. another state of energy defi- journal. ARTICLE IN PRESS 1150 P. but no effects on the anorexic behavior were 2003).. 2005).. gonadal or some other endocrine dysfunction that predisposes them for the disorder. menstruation in starvation is not only a normal but If endocrine changes are epiphenomena to a necessary physiological response to the shortage starvation. Such treatment can be danger- Another one would be redundant. hypothalamus and the pituitary in the same axis toms of anorexia nervosa are physiological adapta. 2005). An inhibitor of food intake cannot increase in the it is long known that hormonal changes in anorexia blood as an individual eats more food. 2005). 2006).. On occasion..g. high levels of leptin (Haas et al. ciency (Chan and Manzoros. Human biology has evolved endocrine changes which have been described in to adapt to recurrent periods of famine and patients with anorexia nervosa. Thus. This is leptin affect behavior or if it is the other way not surprising.g. However. 1986). While there may be endocrine changes that are 3. eating disorder symptoms women who develop anorexia nervosa have an dissolve. Once eating (Nillius et al. Obviously. adipose. pituitary. 4. GnRH mediates between the Michael’s approach and suggest that most symp. no evidence that the disorder is associated with or inhibitor of food intake. This work is descriptive. are beneficial tal. As an example. historically. the nervosa ‘‘have little specificity for this disease and observation suggests that. individuals who have managed to meet this hormonal therapy (estrogen. it is not surprising that treatment with of food. Originally thought to mediate not because there is something wrong with starving negative feedback from adipose tissue and prevent humans but because parenteral nutrition is wrong overeating. suggested that leptin does assume a role as an early studies on the functional changes in the inhibitor of intake after the initial period of hypothalamic–pituitary–adrenal axis of anorexic enhanced secretion and that the ensuing elevated patients showed that these were reversed upon levels of leptin not only cause an arrest of weight weight restoration (Gold et al. Similarly. Rather.. anorexia is ‘‘a low leptin state’’ (Chan and Manzoros. can activate ovulation and normal tions to starvation and that treatment should aim menstrual cycles in patients with anorexia nervosa at restoring a normal eating behavior. different role in signaling energy shortage. mood and behaviour’’. possibly together with hormones on affect. 2005) and leptin is now examined in the context of anorexia nervosa. rather than experimen- perhaps including anorexia nervosa. More gain but also predispose anorexic patients for interesting. the case of leptin anorexic patients. The neuroendocrinology of unfavorable to weight gain during treatment of starvation. hypothalamic. 1982). 2006). 2005). in dealing ‘‘directly with the effects of merely suggests that leptin. even. caused by neural or non-neural damage. as was originally proposed. because. tion in starvation. starvation around. e. Orexigenic and anorexigenic peptides as anorexic patients start to eat more food and gain weight. other hormones. We will follow a similar manner. the rate of leptin secretion increases (Haas Anorexic patients eat too little food and there are et al. 1981). as ovulation in the ‘‘low leptin state’’ of hypotha- When Richard Michael (1975) launched this lamic amenorrhea. DHEA) has been used challenge have been at an advantage (Diamond. if anything. it has been starvation’’ (Fichter et al. even lethal (Weinsier and Krumdieck. leptin is now believed to have a rather in this condition. these may have been induced by Many reviews of the neuroendocrinology of eating improper treatment rather than by a factor that is behavior and body weight regulation have been related to the patient. 2005).. Let’s instead ous. Like other forms of starvation. and. thus. leptin is not merely an no reports that reversal of the endocrine conse- . not a sign of dysfunction or disease leptin can reactivate a physiological function such (Södersten et al. he predicted that psychoneuroendocrinol. and we do not know whether the changes in to the individual (Chan and Manzoros. Södersten et al. briefly consider leptin. endocrine changes in states of energy deficiency. The situation is similar for most of the is a common condition. Kishi and Elmquist. There is no evidence that behavior has normalized. cessation of ovulation and reported (Fisher... In physiology of man’s emotional life’’.

the is homeostatically regulated. On this perspective. A quick change for the regulation. 2006). an anorexigen. the food? The common answer is that their brain is supporting evidence is weak (Södersten et al. Third. We also suggest (Södersten et al. If leptin is (reviewed by Södersten et al.. in guiding important to administer large quantities of food. case of OCD lates that peptidergic neural pathways keep body weight constant by stimulating and inhibiting There is an abundance of reports that starvation eating behavior in response to signals from periph. 1975). there is evidence Johnson. neural and obese in reducing the synthesis of NPY in the biochemical function (van Swinderen and Green- hypothalamus (Goldstone et al. This is not surprising considering that down-regulating the production of orexigenic pep. 1945). anorexia nervosa is there are no orexigenic or anorexigenic peptides. NPY. genetically determined anxiety disorder which is 2000) and neuropeptide Y (NPY) (Nergårdh et al. also causes make the obese eat less food. it is easy to understand why some- static regulation of body weight may be a labora. e. first expressed prepubertally as an OCD. 2003) around puberty as anorexia nervosa and subse- can both stimulate and inhibit food intake and the quently. A trait-related permanent change in been defined. Why would someone constant in human evolution but has more likely who has an OCD eat less food? fluctuated (Stubbs and Tolkamp.. knowledge imposed constraints in the availability of food. on the OCD hypothesis. However. of the neurobiology of OCD does not help us body weight has most likely not been maintained understand anorexia nervosa. 2002). This time in the between the ‘‘anxiety’’ genes and behavior context of obesity. However influen.. thought of as a mental disorder that is caused by Neural pathways that are orexigenic under some another mental disorder.. We suggest that this cognitive change. more often the oppo- is because the theory is overly simplistic and that site is hypothesized. Nor has it generated anything that can be used to in addition to endocrine change. First. genetic networks interact in producing variable tides. As pointed out above. 2005). a ‘‘high leptin state’’. What about the changes in orexigenic and anorexigenic peptides that occur in starvation? The fact that anorexic patients do not eat enough food violates the theory that body weight 5. But this is not the case. as a variety of anxiety physiological contexts in which they do have not disorders. et al.. Thus. that leptin stimulates food intake in obesity. and eventually most of the distribution of resources rather than by cycles of time is spent thinking about food and searching for . causes reversible psychological change.g. Because of externally stood (Chamberlain et al.. 2005). effective in treating anorexia as well.. In these cases it was tial this theory has been. 2005).. 2006). Second. Thus. 1997).. but it is not mentation (Ruffin et al. and still is. This proves that eating too little food. 2006). if research on the neuroendocrinology of body weight necessary under compulsion. High leptin phenotypic expression patterns (pleiotropy). humans have effective also in OCD. why do the obese eat too much The OCD hypothesis is problematic. probably because the neural evolved with unpredictable periods of famine and systems engaged in OCD are only partially under- intervening periods of feast. it has not yet yielded anything that can better was observed when this was done’’ (Burger be used to make anorexic patients eat more food.. To the contrary. one who is starving would develop an OCD. then 2006) as well as CRH (Samarghandian et al.. The psychology of starvation. 2004). et al.. while treatment with GnRH surprising that most of the presently used pharma- activates ovulation and menstrual cyclicity it does cological methods of weight control have minor not affect eating behavior and weight (Nillius effects. By contrast. parti- levels have the expected effect on the brain of the cularly those involved in behavioral. Homeo. insensitive to all the leptin and does not respond by 2006). Most recently. Short- tory artifact and it is more likely that eating age of food gradually increases the time spent behavior is normally controlled by ‘‘discontinuous thinking about food. leptin (Ammar et al. it has been circumstances can be anorexigenic under other hypothesized that the cause of anorexia is a circumstances. a pharmacological treatment of OCD should be possibility which is supported by animal experi. neural serotonergic function is thought to mediate Consider leptin once more. Pharmacological treat- that body weight is not normally homeostatically ment (most often SSRI) is in fact marginally regulated. We suggest span. ARTICLE IN PRESS Psychoneuroendocrinology of anorexia nervosa 1151 quences of their starved condition affects their physiological depletion and repletion’’ (Collier and eating behavior. For example. it is not that it does not. To cite just eral energy storing organs and the gastrointestinal one: ‘‘Psychosis secondary to starvation was evi- tract (Kishi and Elmquist. dent to a variable degree. later in life. This theory postu.

50 the question arises whether the ‘‘OCD of anorexia 16 25 nervosa’’ is an OCD.7 months with this method. ARTICLE IN PRESS 1152 P. Saito. Thus... they reached an average BMI of 17.7 years old 2006). 1877. however. on Michael’s (1975) physio.470. Supply of warmth is useful in reducing the treatment in which re-learning how to eat is activity and facilitating food intake (Gull. We suggest that it is not. example. F. 2005). These observations suggest that there are no interpretation is. Södersten et al. The Bergh et al. For Ammar. kg/m2) in 40. 1874. T.. an 10 ‘‘obsessive’’ trait was recently suggested to be 0 20 40 60 80 100 asset in other circumstances as well: in scientific Time to remission (%) work (Lawrence. A. Time from admission to remission is expressed as percent. A. In fact. food (Keys et al. (mean7SD) anorexic patients. We suggest that any other treatment has been evaluated in a randomized intervention aimed at reducing the physical hyper- controlled trial and found to bring about 75% of the activity of anorexia nervosa should be researched. who were treated to remission in 17. we have developed a 1874). Just 0 as the cessation of ovulation is a physiological 13 response... analysis of such disorders in the context of anorexia nervosa has not yielded treatments that have improved outcome. circumstances and so the physiological significance Rather than questioning the existence of OCD as a of the results reported in Fig. patients may Bergh et al. A. that there is some. outcome has remained poor during at least 50 years (Steinhausen. we would like to underline that the determined.. we have patients with anorexia nervosa. with a starved individual. Final remark this field. (2002). serotonin turnover or OCD). mass index (BMI..4 in this time but all patients eventually here. that their activity is difficult to control (Gull. There are no reports of patients who are free of symptoms upon discharge from a clinical program. Sederholm. A second. 2006)! Figure 1 Recovery of body mass index (BMI) in 40 patients with anorexia nervosa. This appears to be not only 22 Percentile: a natural. of this particular BMI in about 2–3 months has been A common interpretation of this is that patients associated with high levels of leptin. the obvious aim of research in 7. The low BMI values are 6.R. As we have argued 17. Attainment effects on appetitive and consummatory ingestive behavior . 1 reached a BMI of prone to develop anorexia. but a necessary adaptation to the 100 shortage of food.. Interestingly. important along with other interventions.E... P. 2002).g. leptin may do just that (Chan and For a detailed description of this treatment see Manzoros.A. Scheurink. Interestingly. 1950). To improve this situation. However. 2000. starvation is a common state. experience relief from some of the psychological Figure 1 shows the recovery of a normal body effects of starvation as an extra benefit (Lu et al. 2002). if it does. high neural weight gain and risk for relapse (Haas et al. Treatment from patients who were only 12 years old. more likely 2002).J.4 Johnson. And when patients been unable to find any information on how women are discharged from such treatments they are at normally recover from starvation under other best in partial remission and at risk of relapse. which makes them Several of the patients in Fig.. 2005) and.. psychoneuroendocrine constraints for anorexic pa- thing wrong with the treatments that are offered to tients to regain a normal weight. 19 75 BMI (kg/m2) logical framework for psychoneuroendocrinology. However. patients into remission in on average 14 months. Södersten. so is thinking of food in anorexia nervosa an adaptation to starvation.679. NPY-leptin: opposing in 6 (2–30) months (median and range). There are very few exclusion criteria for partici- pating in this treatment and so the patients reach References their target BMI in somewhat variable times. and it regained a normal BMI and fewer than 10% relapse is unlikely that there is something inherently wrong during a five year follow-up program (Bergh et al. arrest of body have a trait-related disorder (e. 1 remains to be mental disorder. It is long known that patients with anorexia nervosa neither have neuroendocrine studies improved the are hypothermic and physically hyperactive and outcome of treatment of anorexia. therefore.

Gwirtsman. H. 2002.. Psychiatry Bergh. Metab.H... 2006.L. C. Muller. P.. Neurosci.A. H.. Strubbe. 1975. Genetics 169. Zhang.. The overzealous total parenteral nutrition: the refeeding syn- University of Minnesota Press. 282–283. G. 10. S. compulsive disorder: the importance of failures in cognitive A. Behavior.. K.S. R.. Regul.... H.... 1.-C.. Frazer. W. Comp. M. M. P. Neuroscience 116.J.F.. Death resulting from Taylor. J. Clin. R. B. K.. Yamauchi. The neuropsychology of obsessive Nergårdh. 95.. 429–444.. Gynecol.. 1986. Nillius... Södersten.006. press. R1627–R1633. Haas. A. drome revisited. A. B. Acad. T. 81. J. bulimia nervosa. Samarghandian. Ebert. Physiol.1002/erv.. Michael. 1874. eating disorders. Wide. Nieman.. D... 2005. J. Disord. C.. Appetite 29. 50. Am. Kim.. Soc.. J. Drummond. experimenter control. Ritt.C. 1997. 286. are involved in tail-pinch-induced food intake of rats. Abnormal Steinhausen. Am.T.. human illness but not in Prader–Willi syndrome and other 657–676. L. A. W. Trans. T. Olmsted. J. W.. Walsh. Lindberg.. adolescents. A. Henschel. L. Pirke.. R.. 1982. Kave. Manzoros. H. obese subjects... 2005. Attia. Brodin. J.C. women.E.. Biol. Rev.)... Sci. J. and body weight regulation in patients with anorexia nervosa C. 2006. Kuipers.. T.Y. infusion in anorexia nervosa. E. Bosy-Westphal... Brabant.. 2005. The outcome of anorexia nervosa in hypothalamic-pituitary-adrenal function in anorexia nervosa. Scheurink. USA 99. Lancet 366. J. A.J. Am. Burger.M. A.W. 1593–1598. Pediatr. USA treatment of anorexia nervosa? Eur. 2005..P. V. Physiol. Role of leptin in energy.R. Onur.. 2151–2163.750.. van Dijk. M.. Nature 423. G.W. issue. 2003. PLoS. Blackwell. Woodside. 5–16.. Med. Acad. in anorexia nervosa: a randomized controlled trial.K. P. 13–15. Michelsen. Collier. Chan. P.. Zandian.. Södersten.. Rockert. Pike. Leon. Adage.. The Biology of Human Starvation. C. J. 2005. Jimmerson.2006.. Shibasaki. Rev. Starvation Mondelli. 314.. 1981. 4. Chrousos. Integr. Psychoneuroendocrinology. J. J. submitted for markers. Lu.. Parides. R. Fries. D. Doerr.08. 122. Kishi. Acta Psychiatr. 2006. 1284–1293. Tolkamp. publication. Psychiatry 159.. Lancet 1.J. X. Leptin Carter. 74–85. Natl. 29. C.D. H. Engl. 1335–1342. N.A. G. ARTICLE IN PRESS Psychoneuroendocrinology of anorexia nervosa 1153 and sexual behavior. J. doi:10.R. U. and ghosts in science. A. Clin.. S. Swaab..P. Longstreet Weinsier. 2002. The double puzzle of diabetes. Unmehopa. C.L. Feeding and temperature responses Diamond.. anorexia network affecting a simple behavior in Drosophila melano- hysterica). gaster. 519–524. van Swinderen. 1945... Lund. 9486–9491.. Neuroendocrine effects of citalopram in western Holland.J.. 132–146. Behav. Control of energy balance Goldstone. D. G.. H.. Psychoneuroendocrinology 1. Fluoxetine after weight restoration before and during weight recovery. R.. J. R. C.P. 2006. 393–399. and young adults. Scand. Clin.C. B. nervosa. Psychoneuroendocrinology. Paul. U. 2006. S. Eat..M. S... 399–419. Osgood.. M. P.A. Devlin. 87. Flexibility in a gene Gull.L.. A. Mol.. master. V. N. Nutr. 2006.. W. A. Nutr. D. 927–937. Ammar. Corticotropin-releasing factor as well as opioid and dopamine Fisher. G. 921–928. Scheurink. in relation to energy intake and energy expenditure in Hypothalamic NPY and agouti-related protein are increased in animals and man: an ecological perspective. Comp. Am. Randomized controlled trial of a treatment for anorexia and Lawrence. B. P.. J.. O.L.. 66. 2002.. . M.. London 7.. the 20th century. doi:10. 1950. Loriaux. Kaplan. Br. Body weight is regulated by the Physiol. 572–578. Sandstead. M. Fineberg. nutrition and endocrinology in anorexia nervosa. Gold.. Understanding Gallucci. B. 159–180. D.. Brodin. Ohata. Physiol.1016/j. to intravenous leptin infusion are differential predictors of 599–602. Bergh... Sci.. Bergh.. 2006. Keys..R.... Avgerinos. I. Krumdieck.O. 2003.. T. J.. 1975. 2605–2612. Horm. Men.T. 278. Am. (Ed. S. W. brain: a link between feeding and emotion. Maletz. 34. J. Fichter..P. Am.. Sahakian. R756–R763. Who is in charge? Animal vs.J..K.J. 2004.. Alters. Clin. Regul. J.J. Proc.M. Successful induction of deprivation states: normal human physiology and clinical follicular maturation and ovulation by prolonged treatment implications for hypothalamic amenorrhoea and anorexia with LH-releasing hormone in women with anorexia nervosa. R. P. Proc. JAMA 295. this Chamberlain.S. Nutr.. Hauer. M. Leptin: a Södersten..S. P. Obstet. 22–28.. D. Södersten. obesity in rats.. Stubbs.J. How effective is family therapy for the potential novel antidepressant... L. Biobehav. attitude. Elmquist. Klein. N. G. M. 2006. Neuropeptide Y facilitates activity- and behavioural inhibition as candidate endophenotypic based-anorexia.. Integr. M. Anorexia nervosa (apepsia hysterica.. Ruffin.W. 889–896.. Editorial. Natl. Bergström. Rev. Greenspan. M. M. Roberto. Minneapolis. 27. Johnson. Bloom.. Robbins. Treatment of eating disorders in children.F. F. Brozek. in 103.psyneuen. U. H. T... 2006. Endocrinol. Nutzinger.