Investigation and management of moderate to severe inpatient hyponatraemia in an Australian tertiary hospital

Hyponatraemia is defined by a serum sodium (Na) concentration <135 mmol/L [1]. It is the most common electrolyte disturbance amongst hospitalised patients at a rate of 15–30% [1‐3]. A prospective cohort study of 98,411 hospitalised patients demonstrated that even mild hyponatraemia (Na 130–134 mmol/L) is associated with increased mortality in hospital (OR 1.37), at 1 year (OR 1.35) and at 5 years post discharge (OR 1.24) [4].

Traditionally, hyponatraemia is classified clinically by volume status into hypovolaemic, euvolaemic and hypervolaemic hyponatraemia [1, 5]. The most common cause of euvolaemic hyponatraemia is the syndrome of inappropriate antidiuretic hormone (SIADH) [3, 6], which is characterized by dysregulated antidiuretic hormone (ADH) secretion and water retention, despite low serum Na and osmolality [5‐7]. Diagnostic criteria include: euvolaemic state, normal thyroid hormone and cortisol levels, serum Na <135 mmol/L, reduced serum osmolality <275 mOsm/kg, urine osmolality >100 mOsm/Kg and urine Na >30 mmol/L [2].

Hyponatraemia can present a diagnostic challenge in patients with cancer [8]. SIADH is the most common aetiology in this setting and may be secondary to lung or brain neoplasm, paraneoplastic syndromes or as a result of chemotherapy agents, NSAIDs or opioids [8]. Other aetiologies to be considered include excessive fluid administration and dehydration [8].

Identifying the correct aetiology of hyponatraemia requires a thorough history, physical examination and consideration of relevant biochemical results. Serum electrolytes (including Na, glucose and urea), serum osmolality, with paired urine Na and osmolality are all necessary to establish the correct diagnosis [1, 2]. In the case of euvolaemic hyponatraemia, adrenal insufficiency and hypothyroidism should also be excluded [2], although in practice hypothyroidism is rarely a cause of hyponatraemia [1].

The treatment of hyponatraemia depends on both the underlying aetiology and the timeframe in which it develops [1, 2, 8]. The recommendation by both the American expert panel and European clinical practice guidelines for first line treatment of chronic euvolaemic hyponatraemia, with either mild or absent symptoms is fluid restriction; however, the paucity of supportive data for this is acknowledged [1, 2]. There is no clear guidance regarding the volume to which fluid intake should be restricted. Some groups advocate the use of the urine/plasma electrolyte ratio as a predictor of non-response to fluid restriction due to negative free water clearance if >1 [9, 10], while it has been previously reported that a urine osmolality of >400 mOsm/kg correlated with failure of fluid restriction [11].

The underlying aetiology and outcomes of hyponatraemia have been examined previously in Australian studies [12, 13], however the investigation, accuracy of diagnosis and detailed management practices have never been described in the Australian setting. The aim of this study is to examine the aetiology, investigation and treatment of moderate to severe hyponatraemia in an Australian tertiary hospital.

Case finding was undertaken using the pathology laboratory program AUSLAB at the Princess Alexandra Hospital, an adult tertiary referral centre with >100,000 admissions per year. Patients who were admitted from the 1st of March to 31st of May 2016, and had moderate to severe hyponatraemia, defined as a serum Na ≤125 mmol/L at any point during an admission were identified. This degree of hyponatraemia was chosen as clinically significant, because serum Na >125 mmol/L has been correlated with an absence of symptoms in prior studies [14, 15]. Medical records were reviewed by three investigators (KB, KH, LL). Exclusion criteria included age <18 years, pregnancy and pseudohyponatraemia secondary to hyperglycaemia or hyperlipidaemia.

Parameters recorded included: patient demographics, medications, serum and urine electrolytes, thyroid and adrenal function. Duration of hyponatraemia was determined from the available blood test results and from the clinical record. If the duration was unable to determined it was classified as unknown. Volume status was assessed by treating team documentation (noted as hypo-, eu- or hypervolaemic) or by documentation of clinical features such as capillary refill, pulse rate, blood pressure (including postural blood pressure), jugular venous pressure, mucous membranes and presence of pulmonary or peripheral oedema. Symptoms of hyponatraemia such as headache, nausea and vomiting, confusion, seizures and coma were noted [5, 6]. The treating team’s working diagnosis (if documented) was also recorded. Patients were analysed and re-classified into likely underlying cause of hyponatraemia if this could be determined from the documentation of fluid status as well as available investigation results. A diagnosis of SIADH was considered probable if the criteria presented by Spasovski et al. were met [2]. A diagnosis of non-renal salt depletion was deemed likely if the urinary Na was <30 mmol/L in the setting of hypovolaemia. A diagnosis of fluid overload was deemed likely if the urinary Na was <30 mmol/L in conjunction with clinical features of hypervolaemia such as oedema, ascites or pleural effusions [2]. Patient diuretic use was assessed to be a significant contribution to hyponatraemia if the urine Na was ≥30 mmol/L [2]. Primary polydipsia causing relative water excess was diagnosed if the urine osmolality was <100 mOsm/kg [2]. Information was also collected regarding management, change in serum Na (Δ Na) and outcomes.

There was a total of 9774 admissions (excluding day admissions) over the three-month period. One hundred fifty-two inpatients (1.6% of admissions) were identified with a serum Na ≤125 mmol/L with a median age of 66 years (IQR: 54–75 years). Patients were predominantly admitted under surgical units (41 of 152 patients (27%), General Medicine (38 patients (25%)); followed by Cardiology (16 patients (10.5%)), Nephrology (13 patients (8.6%)), Gastroenterology (13 patients (8.6%)) and Respiratory (7 patients (4.6%)). Table 1 summarizes the patient demographics, comorbid conditions, contributory medications, symptoms and estimated duration of hyponatraemia.

In 106 patients (69.7%), there were no documented symptoms associated with hyponatraemia. The most prevalent symptom recorded was nausea and/or vomiting in 21/152 patients (13.8%). There was a higher incidence of reported symptoms in patients with more severe hyponatraemia, with symptoms reported in 20/30 patients with a Na level <120 mmol/L, compared with only 26/122 patients of patients with a serum Na 120–125 mmol/L, chi square analysis P <0.01. The median (IQR) nadir serum Na in the symptomatic group was 121 mmol/L (115–124) compared with 124 mmol/L (122–125) in the asymptomatic group, Mann Whitney U - P <0.01).

This study of hospitalized inpatients with moderate to severe hyponatraemia is the first to examine the investigation and management of this condition in the Australian setting. Our data suggest that inpatient hyponatraemia is often inadequately investigated, leading to errors in diagnosis and treatment inconsistencies.

Investigation of hyponatraemia was often incomplete, with less than half of the cohort having urine Na and osmolality performed. Similarly, thyroid function tests and morning cortisol levels were often omitted in euvolaemic patients. Comparable data have been reported by other studies conducted internationally, with adequate investigation of hyponatraemia occurring in only 26–47% of patients [16‐18]. Likewise, Huda et al., found only 27% of patients had a urine osmolality and only 10% had urine sodium performed [16].

In this cohort, euvolaemic hyponatraemia was most frequently described. Other studies have found that SIADH was the most common cause of hyponatraemia [3, 16] but following adjudication by the investigators, only 21.1% were judged likely to have SIADH. As 31.6% of patients did not have a diagnosis documented and 19.7% of patients had inadequate investigation to retrospectively determine a cause of hyponatraemia, it is likely that many undiagnosed patients had SIADH to account for this discrepancy. Diagnostic precision was also examined by Huda et al., who found the inpatient diagnosis was inconsistent with the available clinical and investigative data in 42% [16].

Within the SIADH cohort, fluid restriction was first line treatment in the majority of patients, however the degree of fluid restriction appeared to be arbitrary. Urine output for the preceding 24 h was not readily available. In no case was a calculated urine/plasma electrolyte ratio utilised to determine adequate fluid restriction volumes. Patients with a urine osmolality greater or less than 400 mOsm/kg did not have a significant different Δ Na over 3 days of fluid restriction, which may be due to the small number of patients analysed. Other groups have demonstrated that a higher urine osmolality correlated with non-response to fluid restriction [11]. Winzeler et al. prospectively analysed 82 patients with SIADH and treated with fluid restriction of <1000 mL. They found a higher median urine osmolality of 432 mOsm/kg (IQR 331–597) in non-responders compared with 385 mOsm/kg (IQR 301–438) in responders (P = 0.03) [11].

Second line treatments were only used in a small proportion of patients. Of the patients who received urea there was a substantial increase in serum Na levels over the following 72 h, noting that this group had either already failed fluid restriction or were unable to be fluid restricted. Adverse effects, while common, were mild and easily tolerable. These pilot data resulted in a change of practice at our institution, such that urea is now routinely used in patients with SIADH who do not respond to fluid restriction within 24-48 h.

Previous research which supports the use of urea for the management of SIADH includes a retrospective analysis of 42 patients who were admitted to the Intensive care unit [19]. Patients were treated with a standardised protocol which implemented urea after clinical deterioration (measured as drop in GCS by 2 points) or serum Na <130 mmol/L despite 0.9% saline. Hyponatraemia was corrected in all patients and was well tolerated with no adverse effects [19]. Similar data were obtained by Decaux et al. who compared 50 patients with mild hyponatraemia to 35 patients with severe hyponatraemia treated with urea. All the patients in the mild and severe group achieved resolution of hyponatraemia, however 6 patients in the mild group developed hypernatraemia without any associated adverse outcomes [20]. The only prospective data to date compared 1 year of treatment with a vaptan followed by 1 year of treatment with urea in 13 patients. In this cohort, urea was shown to have the same efficacy as vaptans [21].

Hypovolaemic patients treated with 0.9% saline demonstrated a significant increase in serum Na, suggesting that treatment of non-renal salt wasting is less refractory than SIADH. Hypertonic (3%) saline was utilized in cases of severe symptomatic hyponatraemia using varying volumes and over different time periods. There was a significant increase in serum Na using this mode of treatment, however the time frame in which this occurred varied between patients. Frequent monitoring of serum Na is necessary to prevent over correction.

While acknowledging the limitations of the retrospective design, this is the first Australian study to examine the investigation and management of moderate to severe inpatient hyponatraemia in detail, including the efficacy of fluid restriction and in particular the use of second line treatments. The data have been analysed according to documentation by the treating team, as well as reanalysed by the investigators. It is possible that volume status and diagnosis were discussed but not recorded in patients’ notes, however laboratory data has been captured for all analysed patients.

Hyponatraemia is often inadequately investigated, which may lead to diagnostic and management errors. There is a need for improved guidance to clinicians with respect to the recommended initial fluid restriction volume. For patients who do not respond to fluid restriction within 24–48 h or who are unable to be fluid restricted, treatment with urea resulted in improvement in serum Na. Hypertonic (3%) saline remains the treatment of choice for severe symptomatic hyponatraemia but must be carefully monitored.