Received: 29 January 2019 Revised: 6 April 2019 Accepted: 15 April 2019

DOI: 10.1002/eat.23087

CLINICAL CASE REPORT

Central diabetes insipidus associated with refeeding

in anorexia nervosa: A case report

Elaine L. Rosen MD1 | Apisadaporn Thambundit MD2 | Philip S. Mehler MD3,4 |

Steven D. Mittelman MD, PhD2

1
Division of Adolescent Medicine, Department
of Pediatrics, UCLA Children's Discovery and Abstract
Innovation Institute, David Geffen School of Anorexia nervosa (AN) has been associated with a multitude of hypothalamic pituitary
Medicine, UCLA, Los Angeles, California
2 abnormalities, although it is unknown which aberrations reflect disease causation and
Division of Endocrinology, Department of
Pediatrics, UCLA Children's Discovery and which are the consequences of severe malnutrition. Among these endocrinopathies,
Innovation Institute, David Geffen School of
hypothalamic–posterior pituitary aberrations have been described, including disorders
Medicine, UCLA, Los Angeles, California
3
Eating Recovery Center, Denver, Colorado of osmoregulation. We report the case of an adolescent female with a history of severe
4
Department of Medicine, University of AN, restricting subtype, treated aggressively with multiple hospitalizations. During hos-
Colorado, ACUTE at Denver Health, Denver,
pitalization for severe weakness and lethargy, her course of medical stabilization was
Colorado
complicated by significant polyuria, ultimately diagnosed as central diabetes insipidus
Correspondence
(DI). This is the first reported case, to our knowledge, of a severely malnourished ado-
Elaine L. Rosen, Division of Adolescent
Medicine, Department of Pediatrics, UCLA lescent with AN—restricting subtype developing central DI during the refeeding pro-
Children's Discovery and Innovation Institute,
cess for medical stabilization, thus adding to the small body of existing literature on
David Geffen School of Medicine, UCLA, Los
Angeles, California. disordered osmoregulation in this patient population. This case report raises the ques-
Email: elrosen@mednet.ucla.edu
tion as to whether the frequency of central DI during refeeding is greater than that pre-
Section Editor: ••• viously recognized. Additional research should focus on how neuroendocrine
dysregulation of water balance might impact the clinical course of AN and its
treatment.

KEYWORDS
anorexia nervosa, antidiuretic hormone (ADH), arginine vasopressin (AVP), diabetes insipidus,
osmoregulation, refeeding

1 | I N T RO D UC T I O N AN is characterized by global dysregulation in many endocrine hor-

Anorexia nervosa (AN) is characterized by (a) restriction of energy
intake relative to requirements leading to significantly low contextual
body weight; (b) intense fear of weight gain, or behavior that interferes
with weight gain, even though at significantly low weight; and (c) distor-
tions of body or weight, self-perception, or persistent denial of the seri-
ousness of the low body weight (Diagnostic and Statistical Manual of
Mental Disorders (DSM-5); American Psychiatric Association [APA],
2013). It carries the highest mortality rate of any mental illness and is a
common cause of chronic disease in adolescence, with a lifetime preva-
lence of 1.8% (Smink, Hoeken, Oldehinkel, & Hoek, 2014).
mones (Miller, 2011). The excretion of water is controlled by arginine
vasopressin (AVP, also known as antidiuretic hormone), which is
secreted by the posterior pituitary in response to increases in serum
osmolarity, and acts primarily to increase renal water reabsorption. Dia-
betes insipidus (DI) occurs when the pituitary gland cannot release AVP
(central DI) or the kidneys do not respond to it (nephrogenic DI).
Although dysregulation of the AVP axis has been reported in AN
(Bahia, Chu, & Mehler, 2011; Evrard, Cunha, Lambert, & Devuyst, 2004;
Gold, Kaye, Robertson, & Ebert, 1983; Nishita et al., 1989), we report
herein the first description (to our knowledge) of an adolescent with
AN presenting with DI during hospitalization for severe malnutrition.

752 © 2019 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/eat Int J Eat Disord. 2019;52:752–756.
 1098108x, 2019, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/eat.23087 by Cochrane Saudi Arabia, Wiley Online Library on [29/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ROSEN ET AL. 753

2 | P R E S E NT A T I O N

HD 43c

−2,825
1,500
4,325
The patient was an otherwise healthy 18-year-old female with a

–
–
–

–
4-year history of AN, restricting subtype. Her clinical course was nota-
ble for six neuropsychiatric hospitalizations between 14 and 17 years

−1,501
HD 42

1.006
2,150
3,651
(285)
of age. Upon turning 18, she was discharged from inpatient psychiat-

136

–
ric treatment and refused all forms of psychological care, though she
continued to see her adolescent medicine physician and continued
her fluoxetine (60 mg daily), which she had been on for 1.5 years.

345 > 647 > 708

Over the next 7 weeks, she increasingly endorsed passive suicidal
ideation, stating that she would rather die than gain weight. Her

HD 29b

−2,384
1.008e
weight dropped from 36.4 to 29 kg (body mass index (BMI) 14.9–

1,716
4,100
<0.5
142
300
11.8 kg/m2), with most of the weight loss occurring with near-
complete food refusal during the 6th and 7th weeks. Her reported
water intake throughout this time averaged 4 L/day. She was hospital-

−1,290
HD 16
ized (Hospitalization #1) for hyponatremia in the context of complete

1.006
2,460
3,750
140
300
274
food refusal and placed on a psychiatric hold for risk of self-harm.
Upon admission, the history of high water intake, low serum sodium,
and low urine and serum osmolalities was consistent with that of psy-

HD 15

(XX) represents manual calculation of serum osmolality, using the formula: Posm = 2 × [Na] + [glucose]/18 + blood urea nitrogen/2.8.
1.009
2,200
3,000
−800
chogenic polydipsia (Table 1, Hospitalization #1). Hyponatremia was

(294)
143

393
asymptomatic and managed conservatively with fluid restriction. After

Reference ranges: Sodium, serum mmol/L: 135–145; Osmolality, serum Osm/kg: 285–295; and AVP, serum pg/mL: 0.0–6.9.
48 hr, the psychiatric hold was lifted as her condition was not deemed
to be imminently life threatening even though her inpatient food con-

−1,925
HD 14

1.006
1,550
3,475
(297)
sumption averaged 200 kcal/day. Her pattern of high fluid and low
144

238
caloric intake continued following discharge.
Hospitalization #2 began 2 weeks later when she presented to the
Emergency Department with weakness, lethargy, hypoglycemia

−1,990
HD 12

1.008
2,010
4,000
(302)

(30 mg/dL), and bradycardia (45 bpm). She weighed 27.2 kg (BMI
147

11.0 kg/m2) and was admitted to the medical intensive care unit. A
–
Hospitalization 2 (8/31–10/14)

computed tomography scan of head obtained for lethargy showed an

incidental pneumomediastinum, so she was made nil per os for Days
−3,168
6,432f
1.009

9,600
HD 2

(294)

2–5 and started on prophylactic broad-spectrum antibiotics. The

140

pneumomediastinum was not clinically impactful and resolved over the

first week. Early admission blood tests revealed transaminitis, bone mar-
row suppression, hyperferritinemia, and hypercortisolemia—all indica-
(284)d

1.008
2,308
2,600
−292
HD 1
132

tors of physiological stress related to severe malnutrition. She was again

hyponatremic upon admission (Table 1, Hospitalization #2), presumed
secondary to excess water intake, which was quickly corrected. Nour- Includes blood products, fluid boluses, and medications.
Hospitalization

ishment during the first 4 days of Hospitalization #2 was provided by

1 (8/14–8/17)

Lactated Ringers, correlating timewise with the onset of hypokalemia

<1.005

on Day 2 and hypophosphatemia on Day 3. Electrolyte imbalances were

−240
HD 2
127
273

410
650

promptly corrected. Day 2 was additionally notable for the develop-

73

Last full hospital day prior to discharge.

ment of polyuria with concomitant negative fluid balance. She became

First morning void, prior to challenge.
Hospitalization data

hypotensive, for unclear reasons, and required supplemental fluid, albu-

Serum osmolality (mOsm/kg)

min boluses, and low-dose dopamine. An echocardiogram revealed a

Urine osmolality (mOsm/kg)

“Hospital day” = 0700–0700.

small pericardial effusion but normal anatomy and function, and blood
Desmopressin challenge.

cultures were negative.

Serum Na (mEq/dL)

Once weaned from dopamine (Day 7), she was transferred to a

Hospital daya

medical stabilization ward for patients with eating disorders where a

Serum AVP

Output mL
Date 2017

Intake mL
TABLE 1

Urine SG

refeeding protocol was instituted. Calories started at 1,000 kcal/day,

Net mL

advancing by 200 kcal/day up to 3,400 kcal/day, which ultimately

provided the desired rate of weight gain. She remained fully
b

e
a

f

754
compliant with eating food and did not require nutritional supple-
ment replacement.
On the ward, she continued to have a negative fluid balance, with
high volumes of dilute urine (Table 1, Hospitalization #2). Her intake
and output were strictly monitored, per protocol, by the presence of an
in-room observer, denial of access to the bathroom except a 5-min daily
monitored shower, a bedside commode as the only option for voiding
and stooling, and fluid restriction after 10 p.m. These precautions
ensured adherence to water restriction of 2 L/day (with increased
allowances for continued thirst up to 2.75 L/day on 5 separate days),
effectively ruling out continuing primary polydipsia. She reported
nocturia 2–4 times nightly, an increase from her prehospitalization base-
line of 1–2 times nightly. She remained clinically euvolemic, without
lower extremity or dependent edema, throughout. Mild residual hypo-
tension was asymptomatic and gradually improved, without significant
orthostasis.
DI was clinically suspected for the first time on Hospital Day
12, when serum sodium was 147 mmol/L, with 4 L output of dilute
urine (SG 1.008, osmolarity not measured) and a calculated serum
osmolality of 302 mOsm/kg (reference range 285–295 mOsm/kg). On
Hospital Days 16 and 29, the measured serum osmolality was
300 mOsm/kg, with simultaneous urine osmolalities of 274 and
345, indicating inadequate urine concentration in response to hyper-
osmolality. By formal diagnostic criteria, the diagnosis of DI was con-
firmed on Day 16, when the urine osmolality was <300, while the test
on Day 29 could be interpreted as partial DI (Garrahy, Moran, &
Thompson, 2018). The difference between these two could represent a
fluctuating course or laboratory variation around an arbitrary diagnostic
cutoff for urine osmolality of 300 mOsm/kg. Although the sodium
levels were within the normal range at these timepoints (140 and 142),
this is often the case in compensated DI (Garrahy et al., 2018).
Once the diagnosis of DI is established as the cause of polyuria, a
desmopressin challenge is used to determine whether the origin is cen-
tral (hypothalamic) or nephrogenic. Subcutaneous desmopressin of
4 mcg resulted in effective concentration of urine at 30 and 60 min,
confirming the intact renal concentrating ability, and therefore a central
origin of the DI (Table 2). Patient's preinjection AVP level was
undetectable, also consistent with the diagnosis of central DI. Magnetic
resonance imaging showed no pituitary or intracranial abnormality.
Desmopressin acetate treatment was presented as optional and declined
TABLE 2 Desmopressin challenge
Urine osmolality % Change
(mOsm/kg) from baseline
Baseline 345 – over the course of AN presentation and treatment. For both hospitali-
30 min postdesmopressin 647 87.5
60 min postdesmopressin 708 105.2
Urine osmolality at baseline and post-SQ administration of 4 mcg of
desmopressin. A percent change greater than 100% is consistent with
complete central diabetes insipidus.
1098108x, 2019, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/eat.23087 by Cochrane Saudi Arabia, Wiley Online Library on [29/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ROSEN ET AL.
by the patient, and her polyuria gradually, subjectively improved by
2 weeks postdischarge.
3 | DISCUSSION
Disorders of osmoregulation are known to occur with AN. Sodium
dysregulation, primarily hyponatremia, is a common finding in these
patients. In a study of 1,026 consecutive adults admitted to an inpa-
tient and residential eating disorder program (Mehler et al., 2018),
hyponatremia was present upon admission in 16% of patients with
AN-R. Possible reasons include a “reset osmostat” in the setting of
chronically low osmolality, chronic hypovolemia, decreased salt intake
that impairs renal free water excretion, increased water intake
(e.g., primary polydipsia), and use of antidepressants. On the other
hand, hypernatremia was rarely observed at the presentation. Once
refeeding is initiated, careful attention is paid to the electrolyte and
metabolic abnormalities associated with refeeding syndrome (hypo-
glycemia, hypophosphatemia, hypomagnesemia, and hypocalcemia),
yet there is a paucity of information available as to the course of
serum sodium, bicarbonate, and chloride responses to refeeding.
The diagnosis of DI in a patient with AN has only been reported in
the context of pregnancy (Hayashida, Inagaki, & Horiguchi, 2011), which
is an independent risk factor for DI. Older studies observed inappropriate
polyuria in AN (Russell & Bruce, 1966), suggesting that abnormal osmo-
regulation could be due to renal defects (Aperia, Broberger, & Fohlin,
1978) or hypothalamic dysfunction (Mecklenburg, Loriaux, Thompson,
Andersen, & Lipsett, 1974). Hypertonic saline infusion in both acutely
malnourished and weight-restored patients with AN demonstrated sub-
optimal increase of AVP in the cerebrospinal fluid and plasma, or erratic
or osmotically uncontrolled AVP release, which was resolved 6 months
after weight restoration (Gold et al., 1983). Polyuria was evident in all
patients before correction of weight loss and after short-term recovery,
and findings were attributed to hypothalamic dysfunction. A similar
study (Nishita et al., 1989) confirmed the dysregulation between serum
osmolality and plasma AVP levels, though in this case the disordered
osmoregulation persisted after weight restoration. Although antidepres-
sants may contribute to impaired osmoregulation (Spigset & Hedenmalm,
1997), the anorexic state itself independently impairs osmoregulation
beyond the effect of medication alone (Evrard et al., 2004).
The impact of refeeding on AVP levels was explored in rats: rats
being fasted for 48 hr had lower hypothalamic AVP concentrations than
control rats, and refeeding caused a further decrease in hypothalamic
AVP (Burlet, Jhanwar-Uniyal, Chapleur-Chateau, Burlet, & Leibowitz,
1992). This study is consistent with the observed dynamic changes in
AVP physiology associated with refeeding in AN.
Our patient demonstrates two aspects of osmotic dysregulation
zations, she presented with hyponatremia attributed to psychogenic
polydipsia. In Hospitalization #1, the patient was managed with fluid
restriction, to which she exhibited a modest diuresis of hypotonic
urine, followed by resolution of the hyponatremia and overall fluid
balance. Her appropriate urine output argued against syndrome of

ROSEN ET AL.
inappropriate ADH secretion as the etiology of her hyponatremia,
although the observation time and evaluation were limited, and we
cannot exclude contributions from dehydration, inadequate sodium
intake, or disordered osmoregulation related either to her fluoxetine
usage or the underlying mental illness. During Hospitalization #2, hyp-
onatremia resolved quickly, although her fluid balance remained sig-
nificantly negative for several weeks (Table 1, Hospitalization #2),
eventually presenting as DI.
We hypothesize that the process of refeeding in our patient led to
the development of her central DI, substantiated by the timeline and
supportive data in the case–control and rat studies. With the focus in
clinical management on electrolyte and glucose shifts during the
refeeding process, it is possible that a negative fluid balance or
uptrending sodium values (toward the upper limits of normal) in
patients with AN may not be noticed, and could be more common
than previously recognized. Furthermore, abnormal urinary symptoms
appear to be frequent but underreported in adolescents with
AN. Indeed, 62.7% of adolescents with AN reported nocturnal enure-
sis, stress incontinence, or other symptoms, which could be evidence
of DI. Interestingly, of the 17% reporting nocturnal enuresis, none had
reported this to their medical team (Kanbur et al., 2011). Thus,
unrecognized DI could be more common than has been realized.
This case report highlights a clinically significant problem that
should be recognized and addressed if present in patients with AN. In
the inpatient refeeding setting, fluid and calorie intake are controlled
by protocol, and fluids are generally restricted. Therefore, dehydration
may ensue if careful attention is not paid to the fluid balance. Several
aspects of AN could mask the presence of dehydration, including bra-
dycardia and low BUN and creatinine levels. The low urine-specific
gravities observed in DI could lead healthcare providers to assume
adequate hydration. Similarly, resolution of bradycardia is used as a
discharge criterion, while unrecognized dehydration could confound
this recovery indicator. Finally, polyuria and nocturia can be disrup-
tive, particularly to sleep, which may increase psychological stress and
exacerbate anxiety/depression. All of these factors may prolong
weight restoration, and so treatment with Desmopressin acetate
(DDAVP) could be considered on a case-by-case basis in the setting
of partial or complete DI.
The implications of these findings could include both enhanced
understanding of a lesser known neuroendocrine abnormality in the
pathophysiology of AN and potential for clinical intervention. Clini-
cians are advised to screen for abnormal urinary symptoms over the
course of AN, while a negative fluid balance and/or an increased uri-
nation rate with dilute urine should prompt clinicians to consider and
evaluate for DI. Further research is needed to determine (a) the fre-
quency of disordered osmoregulation in this patient population,
(b) factors correlating with these impairments such as duration of
illness, the use of antidepressants, or the differences within the
DSM-5-defined subtypes of AN (restrictive vs. binge purging, which is
present with distinct electrolyte challenges at baseline), and (c) the
role of newer techniques to improve the diagnostic accuracy in the
cases of hypotonic polyuria, such as use of plasma copeptin levels
(Fenske et al., 2018).
1098108x, 2019, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/eat.23087 by Cochrane Saudi Arabia, Wiley Online Library on [29/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
755
CONFLIC T OF INT ER E ST
The authors declare no potential conflict of interest.
Glenn Waller served as action editor of this article.
OR CID
Elaine L. Rosen https://orcid.org/0000-0002-1226-6530
RE FE RE NCE S
American Psychiatric Association. (2013). Diagnostic and statistical manual
of mental disorders: DSM-5. Arlington, VA: American Psychiatric
Association.
Aperia, A., Broberger, O., & Fohlin, L. (1978). Renal function in anorexia
nervosa. Acta Paediatrica, 67(2), 219–224. https://doi.org/10.1111/j.
1651-2227.1978.tb16306.x
Bahia, A., Chu, E. S., & Mehler, P. S. (2011). Polydipsia and hyponatremia
in a woman with anorexia nervosa. International Journal of Eating Disor-
ders., 44(2), 186–188. https://doi.org/10.1002/eat.20792
Burlet, A. J., Jhanwar-Uniyal, M., Chapleur-Chateau, M., Burlet, C. R., &
Leibowitz, S. F. (1992). Effect of food deprivation and refeeding on the
concentration of vasopressin and oxytocin in discrete hypothalamic
sites. Pharmacology Biochemistry and Behavior, 43(3), 897–905. https://
doi.org/10.1016/0091-3057(92)90423-d
Evrard, F., Cunha, M. P., Lambert, M., & Devuyst, O. (2004). Impaired
osmoregulation in anorexia nervosa: A case-control study. Nephrology
Dialysis Transplantation, 19(12), 3034–3039. https://doi.org/10.1093/
ndt/gfh507
Fenske, W., Refardt, J., Chifu, I., Schnyder, I., Winzeler, B., Drummond, J.,
… Christ-Crain, M. (2018). A Copeptin-based approach in the diagnosis
of diabetes insipidus. New England Journal of Medicine, 379(5),
428–439. https://doi.org/10.1056/nejmoa1803760
Garrahy, A., Moran, C., & Thompson, C. J. (2018). Diagnosis and manage-
ment of central diabetes insipidus in adults. Clinical Endocrinology, 90
(1), 23–30. https://doi.org/10.1111/cen.13866
Gold, P. W., Kaye, W., Robertson, G. L., & Ebert, M. (1983). Abnormalities
in plasma and cerebrospinal-fluid arginine vasopressin in patients with
anorexia nervosa. New England Journal of Medicine, 308(19),
1117–1123. https://doi.org/10.1056/nejm198305123081902
Hayashida, M., Inagaki, T., & Horiguchi, J. (2011). Anorexia nervosa and
diabetes insipidus in pregnancy. Progress in Neuro-Psychopharmacology
and Biological Psychiatry, 35(1), 297–298. https://doi.org/10.1016/j.
pnpbp.2010.11.004
Kanbur, N., Pinhas, L., Lorenzo, A., Farhat, W., Licht, C., & Katzman, D. K.
(2011). Nocturnal enuresis in adolescents with anorexia nervosa: Prev-
alence, potential causes, and pathophysiology. International Journal of
Eating Disorders, 44(4), 349–355. https://doi.org/10.1002/eat.20822
Mecklenburg, R. S., Loriaux, D. L., Thompson, R. H., Andersen, A. E., &
Lipsett, M. B. (1974). Hypothalamic dysfunction in patients with
anorexia nervosa. Medicine, 53(2), 147–159. https://doi.org/10.1097/
00005792-197403000-00003
Mehler, P. S., Blalock, D. V., Walden, K., Kaur, S., Mcbride, J.,
Walsh, K., & Watts, J. (2018). Medical findings in 1,026 consecutive
adult inpatient-residential eating disordered patients. International
Journal of Eating Disorders, 51(4), 305–313. https://doi.org/10.1002/
eat.22830
Miller, K. K. (2011). Endocrine dysregulation in anorexia nervosa update.
The Journal of Clinical Endocrinology & Metabolism, 96(10), 2939–2949.
https://doi.org/10.1210/jc.2011-1222
Nishita, J., Ellinwood, E. H., Rockwell, W., Kuhn, C. M., Hoffman, G. W.,
Mccall, W., & Manepalli, J. N. (1989). Abnormalities in the response of
 1098108x, 2019, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/eat.23087 by Cochrane Saudi Arabia, Wiley Online Library on [29/05/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
756 ROSEN ET AL.

plasma arginine vasopressin during hypertonic saline infusion in Organization data base for spontaneous reporting of adverse drug
patients with eating disorders. Biological Psychiatry, 26(1), 73–86. reactions. Pharmacotherapy, 17(2), 348–352. https://doi.org/10.1002/
https://doi.org/10.1016/0006-3223(89)90009-7 j.1875-9114.1997.tb03718.x
Russell, G., & Bruce, J. T. (1966). Impaired water diuresis in patients with
anorexia nervosa. The American Journal of Medicine, 40(1), 38–48.
https://doi.org/10.1016/0002-9343(66)90185-9
Smink, F. R., Hoeken, D. V., Oldehinkel, A. J., & Hoek, H. W. (2014). Preva- How to cite this article: Rosen EL, Thambundit A, Mehler PS,
lence and severity of DSM-5 eating disorders in a community cohort Mittelman SD. Central diabetes insipidus associated with
of adolescents. International Journal of Eating Disorders, 47(6),
refeeding in anorexia nervosa: A case report. Int J Eat Disord.
610–619. https://doi.org/10.1002/eat.22316
Spigset, O., & Hedenmalm, K. (1997). Hyponatremia in relation to treat- 2019;52:752–756. https://doi.org/10.1002/eat.23087
ment with antidepressants: A survey of reports in the World Health