Multifactorial diabetes insipidus during pregnancy: a challenging diagnosis

Background

Diabetes insipidus (DI) is a syndrome characterised by the excretion of abnormally large volumes of dilute urine, frequently associated with increased thirst and polydipsia. Aetiological causes vary; however, they can be divided into four categories, including central (or pituitary) DI, caused by inadequate production and secretion of antidiuretic hormone (arginine-vasopressin, henceforth referred to as AVP), nephrogenic DI caused by renal insensitivity to the antidiuretic effect of AVP, gestational DI caused by increased placental degradation of AVP and primary polydipsia, caused by suppression of AVP secretion by excessive fluid intake,1 sometimes classified as a separate entity.

DI during pregnancy is a rare condition, estimated to affect 4 in every 100 000 pregnancies.2 During pregnancy, particularly in the second and third trimesters, placental trophoblast produces increased levels of vasopressinase, resulting in increased degradation and clearance of AVP.3 Although this is usually a physiological process, it may result in overt disease, especially in patients with predisposing conditions, and may lead to complications such as dehydration and oligohydramnios. A diagnosis of gestational DI warrants an investigation into the underlying cause, which may reveal a pre-existing subclinical DI (central or nephrogenic). Some pregnancy-related factors that may induce gestational diabetes include abnormally high levels of vasopressinase in women with large placental volumes or altered liver function, which may occur in pre-eclampsia, HELLP syndrome or acute fatty liver disease of pregnancy.4

Published associations between an underlying subclinical nephrogenic DI and transient gestational DI are much rarer in literature and may be underdiagnosed.

Case presentation

A 41-year-old woman with HIV infection as well as chronic hepatitis B virus (HBV) infection presented to an appointment with an Infectious Diseases specialist in January 2020. She had a history of irregular adherence to antiretroviral therapy and had been diagnosed with AIDS in 2009, with a previous hospitalisation due to presumed cytomegalovirus (CMV) encephalitis and Kaposi sarcoma in 2014. She denied alcohol intake, smoking and illicit drug use. She had undergone a full-term pregnancy in 2006 with no complications. Family history was irrelevant.

Six months earlier, she had agreed to restart antiretroviral therapy and had been prescribed tenofovir 245 mg/day, emtricitabine 200 mg/day and dolutegravir 50 mg/day, with good viral control and no reported side effects or relevant laboratory alterations but had missed follow-up consultations after 3 months. On her return, the patient revealed that she had abandoned treatment since the previous month and was currently pregnant (estimated 22 weeks of pregnancy time). She denied taking any medications. Moreover, she reported general malaise since the past 3 weeks. Physical examination revealed blood pressure of 87/61 mm Hg, heart rate of 61 bpm, peripheral saturation of 96% and no fever. Skin turgor was decreased. Discourse was poor and disorganised. Neurological examination revealed ataxia and dysmetria in the finger–nose test. The patient was led to the emergency department for immediate evaluation.

Investigations

Laboratory tests revealed slight anaemia (haemoglobin 101 g/L), de novo renal dysfunction (creatinine 118 µmol/L, blood urea nitrogen 19 mmol/L, with a normal glomerular filtration rate 3 months earlier) with hyperosmolar (osmolality 322 mOsmol/kg) hypernatraemia (serum sodium 161 mmol/L), normokalaemia (potassium 4.4 mmol/L), normoglycemia (glucose 4.6 mmol/L) and normocalcemia (calcium 1.26 mmol/L). Liver parameters such as transaminase, bilirubin and gamma glutamyl transferase levels were within normal range; C reactive protein was 261 nmol/L, with no leucocytosis. Urine specific gravity was 1.013. Kidneys appeared globose but otherwise unremarkable on renal ultrasound.

The patient was then admitted to the Department of Infectious Diseases for further investigation. A 24-hour urine test revealed low urinary osmolality of 216 mOsmol/kg. A hormonal panel demonstrated normal values of thyroid-stimulating hormone (1.82 IU/L, free thyroxine 11 pmol/L), adrenocorticotropic hormone (12.7 ng/L) and serum cortisol (502 nmol/L). Copeptin levels were elevated (22.68 pmol/L), in consistency with the elevated serum osmolality.

HIV immunovirological evaluation revealed a viral load of 2 300 000 copies/mL and a CD4 lymphocyte population of 426 cells/uL. HBV viral load was 86 IU/mL and no CMV DNA was detected.

Obstetric ultrasound demonstrated a near-absence of amniotic fluid and a fetus at the ninth percentile with considerable signs of pulmonary hypoplasia.

Cerebral MRI was performed on the 15th day of admission and, despite revealing large confluent areas of white matter hypersignal in periventricular, subcortical and just-cortical areas, predominantly in the frontotemporal region and cerebellum, they were perceived to be overall less extensive when compared with previous examinations from 2004. No pituitary lesions or alterations were recorded.

Cerebrospinal fluid tap was traumatic, with no detectable CMV or human polyomavirus 2 (JC virus) DNA.

Differential diagnosis

A diagnosis of DI must be considered in any patient with hyperosmolar hypernatraemia and low urine osmolality. Establishing its aetiology is essential for optimal treatment and management; in this case, transient gestational DI was the most likely diagnosis considering the chronological association between pregnancy and disease onset. However, the severity of symptoms and degree of foetal underdevelopment mandated an investigation into other possible contributing factors including a previously undiagnosed subclinical form of central or nephrogenic DI.

Central DI is considered to be the most common form in nonpregnant patients; presentations may differ depending on the degree of deficiency of AVP synthesis.5 Although this patient had a history of central nervous system disorder, we could find no evidence of current neurologic opportunistic infection, pituitary injury or space-occupying lesions, which are frequently associated with central DI6; moreover, elevated levels of copeptin, a surrogate marker for AVP concentration, indicated adequate vasopressin production,7 rendering a diagnosis of central DI unlikely.

Consequently, nephrogenic DI caused by antidiuretic hormone resistance seemed more plausible in this context. Associated causes include congenital mutations in aquaporine or AVP genes as well as acquired disorders, mostly secondary to chronic renal injury or medication-induced tubulopathy.8 In this case, adult-age onset of DI seemed to favour the latter hypothesis.

Considering then acquired nephrogenic DI to be the main underlying contributor to the present syndrome, the source of renal dysfunction leading to AVP resistance remained to be established. Although this patient had a normal glomerular filtration rate up until 3 months prior to hospital admission, she had been living with HIV for more than 10 years, a known risk factor for acute and chronic renal disease; additionally, globose kidney appearance on ultrasound supported the possibility of tubulointerstitial HIV-associated nephropathy.9 Tubular dysfunctions have been implicated in the genesis of nephrogenic DI,5 however, we could find no publications linking HIV-associated nephropathy and DI. Drug-induced DI also seemed plausible and a review of literature placed the antiretroviral tenofovir as a likely causative agent.10 11

Treatment

On admittance, volume and sodium were slowly corrected with hypotonic saline solution. Considering the diagnosis of gestational DI, it was decided to attempt a trial of oral desmopressin (0.06 mg/day). Renal function returned to baseline levels on the third day after admission. Due to persistent hypernatraemia, in an attempt to surpass the effects of increased vasopressinase, it was decided to increase desmopressin dosage to 0.06 mg two times per day, which was followed by clinical and biochemical improvements (serum sodium levels 142 mmol/L and serum osmolality 282 mOsmol/kg 1 day later).

Regarding HIV, highly active antiretroviral therapy with tenofovir 245 mg/day, emtricitabine 200 mg/day and dolutegravir 50 mg/day was reintroduced after normalisation of renal function, with no adverse effects.

After correction of hypernatraemia, the patient displayed a marked improvement of psychomotor abilities over the following week, until complete recovery was attained. Nevertheless, amniotic fluid status remained unchanged and foetal condition continued to deteriorate until pregnancy was deemed inviable and was terminated, with the mother’s consent, at 24 weeks.

Desmopressin was discontinued 3 days after termination of pregnancy. On revaluation 5 months after clinical discharge, although sodium concentration remained normal, copeptin levels were still found to be elevated (20.70 pmol/L).