, Volume 10, Issue 1, pp 87–93

Transient gestational diabetes insipidus diagnosed in successive pregnancies: review of pathophysiology, diagnosis, treatment, and management of delivery


  • Ibrahim Kalelioglu
    • Istanbul Faculty of Medicine, Gynecology and ObstetricsIstanbul University
    • Istanbul Faculty of Medicine, Department of Internal Medicine, Division of Endocrinology and MetabolismIstanbul University
    • Prof. Dr. Turan Gunes Cad. Bau San Veliefendi Evleri
  • Alkan Yildirim
    • Istanbul Faculty of Medicine, Gynecology and ObstetricsIstanbul University
  • Tulay Ozkan
    • Istanbul Faculty of Medicine, Department of AnesthesiaIstanbul University
  • Funda Gungor
    • Istanbul Faculty of Medicine, Gynecology and ObstetricsIstanbul University
  • Recep Has
    • Istanbul Faculty of Medicine, Gynecology and ObstetricsIstanbul University
Case Report

DOI: 10.1007/s11102-007-0006-1

Cite this article as:
Kalelioglu, I., Kubat Uzum, A., Yildirim, A. et al. Pituitary (2007) 10: 87. doi:10.1007/s11102-007-0006-1


Gestational diabetes insipidus (GDI) is a rare disorder characterised by polyuria, polydypsia, and excessive thirst usually manifesting in the third trimester of pregnancy. The etiology is thought to depend on excessive vasopressinase activity, a placental enzyme that degrades arginine-vasopressin (AVP), but not 1-deamino-8-d-arginine vasopressin (dDAVP), which is a synthetic form. This is a transient syndrome and may be associated with acute fatty liver of pregnancy and preeclampsia. The use of dDAVP in symptomatic cases has been proven as a safe method for both the mother and the fetus during the pregnancy. We report a case of recurrent gestational diabetes insipidus in successive pregnancies, which responded to dDAVP and subsided after delivery.


Diabetes insipidusDesmopressinVasopressinPregnancy


Gestational diabetes insipidus (GDI) is a rare endocrinopathy complicating pregnancy with an incidence of approximately four in every 100.000 pregnancies [1, 2]. Polyuria, polydypsia, excessive thirst and dehydration are the main features of the disease. The etiology is thought to depend on excessive vasopressinase activity, a placental enzyme that degrades arginine vasopressin (AVP), but not 1-deamino-8-d-arginine vasopressin (dDAVP), which is a synthetic form with a different N-terminal [3].

GDI can be categorized into three groups depending on the response to arginine vasopressin (AVP) and dDAVP: vasopressin-resistant and dDAVP sensitive, vasopressin and dDAVP resistant (nephrogenic), and vasopressin and dDAVP-sensitive (central) [1, 3]. Although there is an increase in AVP levels in pregnancy to maintain sufficient antidiuretic activity, decreased renal effect due to its increased catabolism by placental vasopressinase may result in and is the main cause of GDI [4]. Another factor contributing to the pathophysiology is transient liver dysfunction in which degradation of vasopressinase in liver is decreased, explaining its association with acute fatty liver of pregnancy and HELLP syndrome [5]. There are various case reports and case series considering diabetes insipidus in pregnancy in the literature. We presented another case that had transient gestational diabetes insipidus that was diagnosed in two successive pregnancies and resolved in the puerperium.


A 35-year old para 1 Caucasian patient was referred in the 35th week of gestation with the symptoms of polyuria, polydypsia, inability to tolerate oral intake, weight loss, and fatigue that began in the third trimester and got worse over time. The prenatal course of the pregnancy was uncomplicated until 32 weeks; after that time, her urination and oral intake progressively increased to the degree that when she presented to our institution, she could not tolerate oral intake of water corresponding to her thirst anymore.

Her past medical history was unremarkable except for an appendectomy and an elective cesarean section at term. She had the same symptoms beginning in the third trimester in her previous pregnancy and was diagnosed as gestational diabetes insipidus. No therapy was given and her symptoms resolved in the second week of puerperium. The family history was unremarkable for endocrinopathies and liver diseases.

The general condition of the patient was moderate; her skin turgor was reduced and her mouth was completely dry. Blood pressure was 90/70 mmHg and the pulse rate 88/min. The physical and gynecological examinations were unremarkable. The ultrasonography revealed a 35th week of gestation with normal amniotic fluid index and fetal biometry. Laboratory data on admission were as follows: hematocrit: 35%, platelets: 216,000 μl, glucose: 99 mg/dl, serum sodium: 143 mmol/l (135–146 mmol/l), potassium: 3.4 mg/dl (3.5–5.1 mg/dl) and chloride: 105 mmol/l (95–107 mmol/l). Liver and renal function tests were normal; Aspartate aminotransferase (AST): 23 IU/l (5–42 U/l), Alanine aminotransferase (ALT): 19 IU/l (5–45 U/l), total bilirubin: 0.36 mg/dl (0.20–1.0 mg/dl), serum creatinine: 0.4 mg/dL (0.7–1.4 mg/dl) total protein: 5.3 gr/dl (6.0–8.0 gr/dl), and albumin 3.0 gr/dl (3.2–5.5 gr/dl). According to urinalysis, the specific gravity of the urine was 1,005, pH was 5.0, and no glucosuria or proteinuria was present. Urinary and plasma osmolality were 93 and 300 mOsm/lt, respectively.

The patient was admitted to the hospital for further assessment. Parenteral fluid replacement therapy with 5% Dextrose was begun. The 24-h urinary output and the oral intake were 22 l and 24 l, respectively. Because of the worsening clinical condition of the patient, water deprivation test was not performed before the therapy was begun [6, 7].

Anterior hypophyseal hormone levels were in normal ranges for pregnancy. Free T4 was 17.2 pmol/l, TSH was 1.230 mIU/l [8, 9]. Sella magnetic resonance imaging (MRI) could not be obtained while the patient was anxious regarding the exposure of the fetus to the contrast material given for the MRI scan. For the differential diagnosis of hepatic disorders, a liver ultrasound scan was done, which was normal. On the day after admission, dDAVP 2 μg/ml was administered intravenously. But hyponatremia and pretibial edema developed and the sodium concentration decreased to 121 mmol/l. Hyponatremia detected within 6 h after dDAVP administration. During this time oral intake of water was 2,200 ml and urine output was 650 ml. Hyponatremia resolved spontaneously by decreased oral intake within 24 h. The posttreatment oral intake and urinary output and electrolyte measurements were presented in a table and graph. Urine output, serum sodium concentration and osmolality stabilized at 10th day of oral dDAVP treatment (Table 1, Graph 1).
Table 1

Urine output, oral fluid intake and plasma osmolality before and after DDAVP administration (amosmol/l)


Before DDAVP

First 6 h after parenteral DDAVP

First 24 h after parenteral DDAVP

Oral DDAVP Day 1

Oral DDAVP Day 2

Oral DDAVP Day 7

Oral DDAVP Day 10

Oral water intake (ml)








Urine output(ml)








Plasma Osmolalitya







Urine Osmolalitya







Plasma Na (mEq/l)








Graph 1

Plasma sodium concentrations during dDAVP treatment

Because of the developing signs of water intoxication, parenteral therapy was quitted and dDAVP 0.05 mg tablet was administered orally, twice a day till the 38 weeks of gestation for twenty days when a cesarean delivery was performed with the indication of a previous cesarean section. The fluid balance and the serum electrolytes were monitored daily. The patient responded well to oral administration of dDAVP. The symptoms resolved rapidly and the serum sodium level was normalized. Mean oral intake and urinary output decreased to eight and seven liters per day, respectively (Graph 2). No adverse effects such as symptoms of water intoxication or abnormal uterine contractions were recorded during this treatment regimen.

The planned cesarean section at 38 weeks of gestation was performed under combined spinal-epidural (CSE) anesthesia with oral dDAVP without shifting to IV dDAVP treatment. 500 ml of gelatin solution (Gelofusine®, B.Braun, Melsungen AG) with 15 mg of ephedrine was given intravenously for prehydration and prevention of hypotension. 20 μg of fentanyl and 9 mg of hyperbaric bupivacaine (Marcaine® Spinal heavy 0.5%, Astra Zeneca) injected intrathecally and the anesthetic level reached thoracic fourth (T-4) dermatome. The operation and the anesthesia were uneventful except for two hypotensive episodes, which were treated successful with additional ephedrine and fluid infusion. A female fetus of 3,150 g and 49 cm was delivered with one minute Apgar score of 10 and umbilical artery pH of 7.351. During the whole operative period, additional 10 mg of ephedrine and 2,500 ml of Ringer’s lactate were used. Urine output was 230 ml. Postoperative analgesia was achieved by patient controlled analgesia device (Abbott Pain Management ProviderTM, Abbott Laboratories, North Chicago, USA) containing bupivacaine 0.1% and 2 μg/ml fentanyl (4 ml/dose with a 20 min lockout interval and 4 ml background infusion).

The postoperative course was uneventful. DDAVP treatment was continued at the same dosage and the urinary output on the first postoperative day was 11 l and the oral and parenteral intake were 13 l in total. The urinary output decreased progressively and was seven liters on the fourth postoperative day, when the patient was discharged from the hospital with dDAVP treatment with a dose of 0,01 mg twice a day. This dose was maintained for a 10 day period after discharge. Urinary output decreased to 5 l at 14th day of puerperium. Thereafter dDAVP stopped. Her oral water intake and urine output were within normal limits (<2,500 ml/day) after dDAVP treatment ceased. Sella MR performed at postpartum 6th month revealed no abnormality. But posterior pituitary bright spot could not be observed (Fig. 1).
Graph 2

Alterations of fluid intake and urine output after DDAVP


Gestational diabetes insipidus refers to the state of excessive water intake and hypotonic polyuria. Those manifesting in pregnancy and referred to as GDI may persist thereafter or may be a transient latent form that resolves after delivery. It may be associated with preeclampsia, acute fatty liver of pregnancy, or HELLP syndrome [10, 11, 12, 13, 14, 15].

Many hemodynamic, renal, and endocrine changes take place during pregnancy. Pregnancy is associated with a change in maternal blood volume, osmoregulation, and arginine vasopressin secretion by the posterior pituitary, and therefore a predilection towards diabetes insipidus occurs during pregnancy. Water retention increases whereas plasma osmolality decreases by an average of 10 mmol/kg [6, 16]. Plasma levels of AVP during pregnancy are similar to nonpregnant levels despite three to four-fold increased metabolic clearance. This increased clearance is caused by a placental enzyme called vasopressinase that degrades both AVP and oxytocin. The level of this enzyme increases 1,000-fold during pregnancy and its physiological role is unknown [7, 17]. Rapid degradation of AVP during pregnancy can unmask latent forms of diabetes insipidus and subclinical diabetes insipidus may become symptomatic during pregnancy [18]. Additionally, marked increase in the production of renal prostaglandin E2 and glomerular filtration rate may contribute to increase in urinary output. Clinical presentation is usually acute with symptoms developing in a few days. Additionally, intolerance for oral water intake, nausea, fatigue, weight loss, and reduced skin turgor may be seen. Many patients require parenteral treatment, which usually consists of dextrose and normal saline; but it is also essential to correct any electrolyte disturbances if present. The differential diagnosis of polyuric and polydipsic states in pregnancy are broad and precise diagnosis is difficult. Detailed history is essential in making the differential diagnoses of diabetes insipidus. Primary polydipsia and head trauma should be excluded. Ingestion of drugs should be questioned such as lithium, mannitol, diuretics and anticholinergic drugs.

Water deprivation test, sella MRI and measurements of plasma or urinary AVP are advocated as useful methods in the differential diagnosis. Water deprivation test, which compares urinary osmolality after dehydration with that after AVP administration, is the best diagnostic method among these. But in emergency situations, with acute onset of symptoms, this test may lead to significant dehydration and is not recommended during pregnancy; in these cases, treatment should not be delayed and diagnostic tests may be limited to measurement of urine and plasma osmolality [6]. Measurements of plasma or urinary AVP are expensive and are needed, when osmolality measurements are inconclusive [18, 19]. The practice of relating plasma to urine osmolality is useful; it can quickly differentiate diabetes insipidus from parenteral fluid excess. The urine osmolality is a measure of the concentration of the urine and is determined by AVP concentration. In a hypernatremic patient, a urine osmolality below that of the plasma suggests diabetes insipidus. If simultaneously obtained urine osmolality is low while the plasma osmolality is high, the patient has central or nephrogenic diabetes insipidus. Nephrogenic diabetes insipidus can be ruled out if there is a normal response to injected AVP, that is urine osmolality obtained 30–60 min after injection increases [20].

In our patient the clinical situation prevented us to perform water deprivation test. But urine and plasma osmolality values indicated the diagnosis of diabetes insipidus. The decrease of oral water intake and urine output after dDAVP confirmed this diagnosis. Initial evaluation includes a complete blood count, serum electrolytes, plasma and urine osmolality, liver enzymes, serum creatinine, bilirubins, and total protein. Physical examination and laboratory tests will be helpful in the differential diagnosis in states of glucosuria, chronic renal disease, sickle cell anemia, hypercalcemia, and hypopotasemia that may present with polyuria. In patients with newly diagnosed central DI, complete evaluation to define the cause of the disease is essential. Sella MRI is useful in excluding lesions of the hypophysis and hypothalamus. Long-term follow-up is indicated, as hypothalamic dysfunction may appear with time [20,21,22].

Our patient had normal oral water intake and urine output after delivery. Although MRI did not show the posterior bright spot indicating intact posterior lobe; the absence of bright spot has been reported to occur in some normal subjects [23] (Fig. 1). Thickness of pituitary stalk was normal as shown in Fig. 2. MRI was normal except for the absence of posterior bright spot. We can not exclude the transient infindibuloneurohypophysitis during pregnancy. Because we could not obtain MRI during the acute phase. But anterior pituitary functions remained intact in the postpartum period. Lymphocytic infindibuloneurohypophysitis is a common cause of what previously considered to be idiopatic DI [24].
Fig. 1

MR imaging obtained six months after delivery showed the lack of bright spot of the posterior pituitary on T1-weighted sequences

Fig. 2

Coronal MRI scan demonstrates that thickness of pituitary stalk is normal and at midline

GDI during pregnancy or immediately in the postpartum period may be associated with preeclampsia, acute fatty liver of pregnancy, or HELLP syndrome [5, 10, 11, 12, 13]. This form usually responds to dDAVP, but not AVP [5, 10]. The possible cause is the inability of the liver to degrade vasopressinase, which leads to increased metabolic degradation of AVP. In rare cases this may present as nephrogenic DI [12, 14, 15]. Therefore evaluation for preeclampsia and HELLP syndrome is indicated in these patients.

Once the diagnosis is established, close monitoring of vital signs, fetal status, fluid balance, body weight, and renal function are essential. Fluid intake must approximate urine volume and other fluid losses. Otherwise hypernatremia and circulatory collapse may develop especially if the patient is adipsic.

The treatment of choice in patients with GDI is dDAVP. AVP is a cysteine aminopeptide whereas dDAVP, l-desamino-8-d-arginine vasopressin has a different N-terminal. Vasopressinase is a cysteine aminopeptidase and cannot degrade dDAVP [25]. It may be used intravenously, intranasally, subcutaneously and recently an oral form has been developed. dDAVP has a prolonged antidiuretic activity; the antidiuretic effects last from 8 to 24 h [26, 27]. In the presence of an upper respiratory tract infection, the amount of absorption decreases. In these cases and in the postoperative period, the parenteral form is used. Lower extremity edema and signs of water intoxication may develop in cases with excessive dosage [7]. In our patient, hyponatremia developed rapidly after parenteral 2 mcg dose of dDAVP. This may be due to the excessive dose of parentaral vasopressin, because oral form was sufficient to maintain water homeostasis. We cannot exclude habitual water intake after parenteral dDAVP in our patient as a cause of hyponatremia. This is also reported to occur after first doses of dDAVP in patients who do not restricted oral water intake. The development of hyponatremia after a single dose of IV dDAVP is striking. Desmopressin carries the potential hazard of dilutional hyponatremia if patient continue to drink inappropriately despite persistent antidiuresis [28].

Experience with the use of dDAVP during pregnancy is increasing; it is safe for both the mother and the fetus [29]. Unlike AVP, it is easy to administer and has prolonged duration of action. Earlier pituitary extracts of AVP contained oxytocin and precipitated uterine contractions and preterm labor. Because even oxytocin-free or synthetic ADH can produce rythmic uterine contractions, synthetic antidiuretic-selective desmopressin (dDAVP) is the drug of choice [30]. Chlorpropamide can stimulate secretion of AVP and potentiate its effects, but it can cause fetal hypoglycemia and neonatal DI; therefore it is not the drug of choice in pregnancy [31].

Special care must be given during delivery when oral intake is limited [32]. For labor and delivery regional analgesia or anesthesia is method of choice [33]. Because of increased urinary output and possible dehydration, attention to the potential hazard of hypovolemia is mandatory during regional techniques. In our case combined spinal-epidural (CSE) anesthesia was chosen for elective cesarean section in order not to limit her oral intake and she could continue he oral dDAVP medication without a need for shifting to intravenous treatment. For two reasons we did not monitor the central venous pressure: First of all, serum electrolyte levels were in normal range and the fluid balance was established and secondly the central venous pressure is only a pressure reading and is not representative of the actual circulating blood volume. Right cardiac function, systemic vascular resistance and posture all affect this measurement. Although there are many case reports describing transient GDI in pregnancy, this is the only one that describes recurrence of this condition in successive pregnancies. The symptoms developed around the same gestational weeks in both situations, at about 32nd week of gestation. In the previous pregnancy, her oral intake was increased to about 20 liters per day and polyuria resolved three weeks after delivery. She did not receive dDAVP treatment. In the succeeding pregnancy, her urine output increased to 30 l per day. Pregnancy is a condition that unmasks the subclinical forms of both nephrogenic and central DI [34]. Because this latent condition is aggravated in severity in the second pregnancy, it might actually be progressive and it may be anticipated that the patient might develop diabetes insipidus in the future, depending on the underlying ethiology, such as lymphocytic neurohypophysitis. Therefore long term follow-up is adviced for our patient.

Copyright information

© Springer Science+Business Media, LLC 2007