Treatment Effect of the SGLT2 Inhibitor Empagliflozin on Chronic Syndrome of Inappropriate Antidiuresis: Results of a Randomized, Double-Blind, Placebo-Controlled, Crossover Trial

Abstract

Significance Statement

The syndrome of inappropriate antidiuresis (SIAD) is a major cause of hypotonic hyponatremia. Despite its prevalence, treatment options are sparse, and data on their effect on hyponatremia-associated morbidity such as neurocognitive impairment are largely lacking. New treatment options are needed. The sodium-glucose cotransporter 2 (SGLT2) inhibitor empagliflozin promotes osmotic diuresis via urinary glucose excretion and could be used as a treatment for chronic SIAD. This randomized, double-blind, placebo-controlled, crossover trial with 14 participants revealed that empagliflozin is well tolerated and effective compared with placebo. In addition, treatment with empagliflozin possibly led to an improvement in neurocognitive function. The results set the stage for further studies evaluating empagliflozin as a treatment option in patients with SIAD-induced hyponatremia.

Background

The syndrome of inappropriate antidiuresis (SIAD) is characterized by a reduction of free water excretion with consecutive hypotonic hyponatremia and is therefore challenging to treat. The sodium-glucose cotransporter 2 (SGLT2) inhibitor empagliflozin promotes osmotic diuresis via urinary glucose excretion, likely leading to increased electrolyte free water clearance.

Methods

In this randomized, double-blind, placebo-controlled, crossover trial, we compared 4-week treatment with empagliflozin 25 mg/d to placebo in outpatients with chronic SIAD-induced hyponatremia. At baseline and after both treatment cycles, patients underwent different assessments including neurocognitive testing (Montreal Cognitive Assessment [MoCA]). The primary end point was the difference in serum sodium levels between treatments.

Results

Fourteen patients, 50% female, with a median age of 72 years (interquartile range [IQR], 65–77), completed the trial. Median serum sodium level at baseline was 131 mmol/L (IQR, 130–132). After treatment with empagliflozin, median serum sodium level rose to 134 mmol/L (IQR, 132–136), whereas no increase was seen with placebo (130 mmol/L; IQR, 128–132), corresponding to a serum sodium increase of 4.1 mmol/L (95% confidence interval [CI], 1.7 to 6.5; P=0.004). Exploratory analyses showed that treatment with empagliflozin led to improved neurocognitive function with an increase of 1.16 (95% CI, 0.05 to 2.26) in the MoCA score. Treatment was well tolerated; no serious adverse events were reported.

Conclusion

The SGLT2 inhibitor empagliflozin is a promising new treatment option for chronic SIAD-induced hyponatremia, possibly improving neurocognitive function. Larger studies are needed to confirm the observed treatment effects.

Clinical Trial registration number:

ClinicalTrials.gov NCT03202667.

The syndrome of inappropriate antidiuresis (SIAD) leads to a reduction of free water excretion with consecutive hypotonic hyponatremia.^1,2 There are diverse causes of SIAD, including various organ dysfunctions, but also stress, such as chronic pain, or secondary to medications.³

Treatment options for SIAD, in addition to treating the underlying cause, are limited.^4,5 The first-line treatment option is fluid restriction,^6–8 which is burdensome and unlikely to be sustained over an extended period, especially in chronic SIAD-induced hyponatremia. Urea effectively treats hyponatremia through the induction of osmotic diuresis.^9,10 Although a recent study showed its safety also over an extended treatment period,¹¹ its poor palatability affects compliance.¹⁰ Vaptans correct hyponatremia through the induction of aquaresis, but they are costly and bear the risk of serum sodium overcorrection.^5,12,13 As a result, treatment often remains inadequate.¹⁴ This is concerning given the reported adverse effects of hyponatremia, particularly neurocognitive deficits and gait disturbances.^15–17 Despite these associations, intervention studies showing reversibility of these symptoms are largely lacking, with only one study indicating improvement of a subdomain of neurocognition upon treatment with vaptans.⁷

We recently reported the efficacy of a 4-day treatment with the sodium-glucose cotransporter 2 (SGLT2) inhibitor empagliflozin in correcting hyponatremia in hospitalized patients with SIAD-induced hyponatremia.¹⁸ The therapeutic effect of empagliflozin is the induction of marked glucosuria leading to osmotic diuresis, likely causing increased electrolyte free water clearance.^19,20 Given its good tolerability and positive effects on cardiovascular and renal outcomes,^21,22 empagliflozin could be an ideal treatment for outpatients with chronic SIAD. However, its efficacy in these patients is unknown.

The aim of this randomized, double-blind, placebo-controlled, crossover trial was to investigate whether 4 weeks of treatment with the SGLT2 inhibitor empagliflozin leads to a greater increase in serum sodium levels than placebo in outpatients with chronic SIAD. As secondary objectives, we investigated whether hyponatremia correction by treatment with empagliflozin leads to improvement in neurocognitive deficits and gait disturbances.

Methods

Trial Design and Participants

This prospective randomized, double-blind, placebo-controlled, crossover trial was performed at the University Hospital Basel, Switzerland from December 2017 to August 2021. The local ethics committee (EKNZ 2017-00701) and the national agency for the authorization and supervision of therapeutic products (Swissmedic 2017DR2127) approved the study protocol and study medication. The trial was registered at ClinicalTrials.gov (NCT03202667). Written informed consent was obtained from all participants.

Eligible patients were 18 years of age or older and had chronic SIAD-induced hyponatremia <135 mmol/L, defined as euvolemia according to clinical assessment, serum osmolality <275 mmol/kg, urine osmolality >100 mmol/L, urine sodium >30 mmol/L, exclusion of hypothyroidism, and hypocortisolism.²³ Patients with acute or transient hyponatremia, severe symptomatic hyponatremia in need of hospital treatment, diabetes mellitus type 1, renal insufficiency (GFR <45 ml/min), heart failure, or known liver cirrhosis or acute hepatic impairment (alanine amino transferase/aspartate amino transferase >3× upper limit);who were pregnant or breastfeeding; or who were under treatment with SGLT2 inhibitors, lithium chloride, urea, or glitazone were excluded.

Trial Procedures and Assessments

Outpatients with hyponatremia due to chronic SIAD from our institution and patients referred from endocrinologists in practice were asked to participate. After fulfilling eligibility criteria and providing informed consent, a medical questionnaire including evaluation of general well-being and possible symptoms and effects of hyponatremia was compiled. Routine physical examination and baseline diagnostics were conducted including blood and urine sampling, which were processed according to standardized operating procedures. Participants then completed the following tests:

• EQ-5D-5L test^24,25: a standardized test assessing quality of life consisting of two parts. The first part was a visual analog scale (VAS) ranging from 0 (=worst health) to 100 (=best health) on which the patients rated their current health. In the second part, the patients were asked to attribute a level (no problems=1 point, slight problems=2 points, moderate problems=3 points, severe problems=4 points, extreme problems/unable to=5 points) to each of the following five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression.

• Montreal Cognitive Assessment (MoCA)²⁶: The MoCA is a 30-point test (≥26 defined as normal) used to assess the cognitive domains visuospatial, executive function, naming, memory, attention, language, abstraction, delayed recall, and orientation (to time and place). The MoCA was chosen because it is a sensitive screening test for the detection of mild cognitive impairment and was superior to the Mini-Mental State Examination.²⁷ A special focus was set on the subtest executive function, ranging from zero to five points.

• Grip strength test: This was measured using a hand dynamometer.²⁸ The best score of three trials of the dominant hand was reported.

• Gait analysis: A quantitative analysis of spatiotemporal gait parameters using the GAITRite electronic walkway system²⁹ was performed at the Basel Mobility Center of the University Department of Geriatric Medicine, Felix Platter Hospital. Gait was assessed from one trial of self-paced, habitual walking and included the variables walking speed, step width, single support time (indirect marker of dynamic balance), and gait regularity (measured by cycle time variability).

Participants were randomly assigned to undergo the first treatment period in the empagliflozin or, with equal chance, the placebo group. Treatment involved one capsule per day (empagliflozin 25 mg or placebo, respectively) for 28 days. Further treatment included limitation of daily fluid intake to ≤1.5 L/d. Participants were asked to maintain their fluid intake during the observation period and to record their daily fluid intake.

After baseline, weekly examinations were performed, including a medical questionnaire, clinical parameters, and blood and urine sampling. At the end of each treatment cycle (i.e., 4 weeks after starting study medication) participants additionally performed the above described quality of life, neurocognition, grip strength, and gait assessments. A washout period of at least 2 weeks between the treatment cycles and a follow-up visit 30 days after completion of the second treatment phase were scheduled.

Trial Outcomes

The primary outcome was the difference in serum sodium levels (in millimoles per liter) after 4 weeks of treatment with empagliflozin 25 mg compared with placebo. Secondary end points included serum sodium levels after 1, 2, and 3 weeks of treatment; derived serum sodium area under the curve (AUC), serum and urinary electrolytes, osmolality, and glucose after 1, 2, 3, and 4 weeks of treatment; course of hyponatremia symptoms as assessed in the medical questionnaire and clinical parameters after 1, 2, 3, and 4 weeks of treatment; and change from baseline to end of treatment in EQ-5D-5L test, MoCA, grip strength test, and gait analysis. In addition, adverse events, defined as any new medical issue or exacerbation of an existing medical issue according to Common Terminology Criteria for Adverse Events v4.0, were recorded. Seriousness and severity of each event was documented and its relation to the study intervention assessed.

Laboratory Measurements

Serum and urine concentrations of sodium, glucose, creatinine, urea, uric acid, and osmolality were measured by the central laboratory of the University Hospital Basel. Serum sodium levels were analyzed by indirect ion selective electrode method (cobas 8000 modular analyzer; Roche Diagnostics). Serum and urinary osmolality were measured using the freezing point depression osmometer method. To ensure the double-blind design of the study, results from the urinary diagnostics after administration of the first study drug were blinded until the end of the study.

Sample Size Estimation

Assuming a baseline value of 127 mmol/L (SD, 3 mmol/L) and a within-patient correlation of ρ=0.70, we estimated that a sample of 13 evaluable patients would provide the trial with 90% power to detect a difference of 3 mmol/L in serum sodium levels after 4 weeks of treatment at a significance level of 5%. The hypothesis test was based on a linear model with treatment as categorical predictor and the baseline value as covariate. Considering a dropout rate of 20%, we planned to recruit three additional patients.

Patient Flow and Analysis Sets

A total of 17 patients were included in the study and randomized (full analysis set) to the sequence empagliflozin–placebo (n=8) or placebo–empagliflozin (n=9) (Figure 1). Of these, three patients were excluded from the analyses: two patients withdrew their consent within/after the first week of treatment during the first treatment phase (one randomized to empagliflozin–placebo and one randomized to placebo–empagliflozin), and one patient (randomized to placebo–empagliflozin) was excluded after the first treatment phase due to initial incorrect diagnosis (transient hyponatremia).

Figure 1.

Study flow diagram.

The remaining 14 patients defined the modified intention-to-treat analysis set (ITTS). All patients in the ITTS took at least one dose of study medication for both empagliflozin and placebo treatments and received the treatments in the sequence as randomized. Of these, one patient (placebo–empagliflozin) took only 18 pills of a total of 28 pills and was not adherent to the treatment as defined per protocol (predefined as 75% of pills taken, i.e., 21 of 28). The remaining 13 patients defined the per protocol analysis set (PPS).

Statistical Analyses

We used linear mixed regression models to analyze the primary outcome serum sodium levels at the end of treatment, with treatment (empagliflozin–placebo) and baseline levels (visit 0) as fixed effects. In order to account for the crossover design, patient was included as random effect (random intercept). Because preceding analyses revealed no evidence for either carryover or sequence effect, treatment sequence and study phase were not included in the primary analysis model (see Supplemental Statistical Analyses).

The primary analysis was performed on the modified ITTS and repeated on the PPS. We performed adjusted analyses by adding serum and urinary sodium and osmolality, and the fractional excretions of urinary sodium, urea, and uric acid as additional covariates (separate models, also including the interaction term covariate × treatment). The adjusted mean treatment effect is presented with 95% confidence interval (CI).

Secondary end points were analyzed analogously, using mixed linear (continuous outcomes) or logistic regression models (binary outcomes). To evaluate the effect of reaching normonatremia (serum sodium level at end of treatment ≥135 mmol/L), separate mixed-effects linear regression models were fitted with the change in MoCA, grip strength, and gait parameters, from baseline to end of treatment as dependent variable, baseline value (covariate) and sodium status at the end of treatment as fixed effects, and patient as random effect.

Secondary analyses were performed on the ITTS. Adverse events are reported for the full analysis set.

Sodium measurements for baseline and week 4 were available for all patients for both treatment phases; however, several patients missed visits in between, due to comorbidities and the ongoing coronavirus disease 2019 (COVID-19) pandemic. To derive the serum sodium AUC, missing values were imputed by multiple imputation using chained equations (see Supplemental Statistical Analyses). For all other secondary end points the analyses were based on the available data.

All analyses were predefined and conducted using the statistical software package R versions 3.6.3 (February 29, 2020) and 4.1.2 (November 1, 2021).³⁰ No adjustment was made for multiple testing.

Results

Baseline Characteristics

Fourteen patients completed both treatment cycles and were included into the analysis (see the study flow chart in Figure 1). Median age was 71.5 years (interquartile range [IQR], 64.5–76.8), with 50% of the participants being female. Main comorbidities were arterial hypertension (79%), and cerebrovascular and psychiatric disorders (both 36%) (Table 1).

Table 1.

Baseline characteristics

Characteristic	Participants (n=14)
Age, yr	71.5 [64.5–76.8]
Sex: female, n (%)	7 (50)
BMI, kg/m2	24.4 [21.6–27.6]
Comorbidities, n (%)
Arterial hypertension	11 (79)
Cerebrovascular disorder	5 (36)
Pulmonary disease	4 (29)
Psychiatric disorder	5 (36)
Diabetes mellitus type 2	2 (14)
Central nervous system disorders	2 (14)
Malignancy	1 (7)
Chronic infectious disorders	1 (7)
Drugs, n (%)
Antihypertensive	9 (64)
Statins	4 (29)
Antiepileptic/multiple sclerosis treatment	4 (29)
Asthma/COPD inhalants	4 (29)
Proton pump inhibitors	4 (29)
Hormonal replacement	4 (29)
Anti-inflammatory	4 (29)
Metformin	3 (21)
Antipsychotic/antidepressants	2 (14)
Diuretics	2 (14)
Pain medication	2 (14)
Antiretroviral	1 (7)
Other	9 (64)
Causes of SIAD, n (%)
Central nervous system disorders	2 (14)
Stress (chronic pain)	1 (7)
Drug-induced	4 (29)
Pulmonary disease	3 (21)
Idiopathic	4 (29)
Duration of hyponatremia, mo	45.5 [15.8–57.3]

Summary statistics of patient characteristics according to the intention-to-treat analysis set. Categorical variables are shown as frequencies (%), numerical variables as median [interquartile range]. BMI, body mass index; COPD, chronic obstructive pulmonary disease.

Median serum sodium level at baseline was 131 mmol/L (IQR, 130–132) (Table 2). The etiology of chronic SIADs ranged from drug-induced (antiepileptic [n=3] and antidepressants [n=1] that could not be stopped) to pulmonary (n=3) or central nervous system disorders (n=2) to stress-induced due to chronic pain (n=1). In four patients, the etiology remained idiopathic. Hyponatremia duration ranged from a minimum of 4 months to a maximum of 90 months.

Table 2.

Summary statistics and treatment effect for clinical parameters, laboratory values, and assessment outcomes

		End of Treatment
Characteristic	Baseline	Placebo	Empagliflozin	Treatment Effect [95% CI]	P Value
Clinical parameters
Body weight, kg	72.0 [69.0–80.4]	73.5 [70.1–79.6]	72.8 [66.4–78.4]	−1.65 [−2.75 to −0.55]	0.009
Systolic BP, mm Hg	149.0 [134.5–160.2]	141 [135.3–152.5]	140.5 [132–146.5]	−5.34 [−14.05 to 3.38]	0.23
Diastolic BP, mm Hg	80.0 [68.8–86.0]	79 [75–93]	80.5 [72.3–88]	−1.74 [−6.94 to 3.47]	0.51
Heart rate, bpm	66.0 [60.8–72.2]	64 [60.3–66]	64.5 [60–68]	0.03 [−6.27 to 6.32]	0.99
Laboratory values
S-sodium, mmol/L	131 [130–132]	130 [127.5–132]	134 [132–136]	4.09 [1.68 to 6.49]	0.004
S-glucose, mmol/L	5.0 [4.8–5.8]	5 [4.73–5.88]	5.1 [4.8–5.8]	−0.41 [−0.81 to −0.02]	0.05
S-creatinine, µmol/L	61.0 [57.5–64.5]	61 [53.3–70.3]	72.5 [59–78.3]	7.76 [3.88 to 11.64]	0.001
GFR, ml/min	87.0 [85.0–101]	88 [82.3–100.5]	81.5 [68.3–95]	−7.09 [−10.97 to −3.21]	0.003
S-urea, mmol/L	3.70a [3.40–4.40]	3.75 [3.3–4.92]	4.75 [3.73–5.68]	0.50 [−0.24 to 1.24]	0.19
S-uric acid, mmol/L	213.0a [160.2–332.8]	200.5 [149.5–304.75]	192 [134.75–271]	−20.09 [−52.39 to 12.21]	0.22
S-osmolality, mosm/kg	271.5 [264.0–281.0]	268.5 [261.75–280.75]	281 [274.5–284.75]	17.16 [0.38 to 33.93]	0.06
U-sodium, mmol/L	97.5 [77.0–130.0]	81.5 [70–101.25]	94 [57.75–107.5]	3.26 [−16.08 to 22.61]	0.73
U-glucose, mmol/L	0.2 [0.2–0.27]	0.25 [0.2–0.38]	131.7 [90.58–155.88]	6.19 [5.79 to 6.59]	<0.001
U-urea, mmol/L	143.0 [109.8–187.8]	155.5 [97.25–208.5]	160 [123–216]	−1.82 [−36.51 to 32.88]	0.92
U-uric acid, µmol/L	2072 [1532–2393]	1989 [1332.25–2583.5]	2009.5 [1578.5–2640.75]	78.31 [−563.97 to 720.59]	0.80
U-osmolality, mosm/kg	468.0 [361.2–567.0]	494 [337.75–592.75]	601 [502.5–676.3]	131.11 [70.61 to 191.62]	<0.001
Assessments
EQ-5D-5L score
EQ-5D-5L (VAS score)	70a [60–85]	75a [60–80]	65a [60–87]	−0.38 [−9.02 to 8.26]	0.93
EQ-5D-5L (unit score)	0.87a [0.72–0.94]	0.8a [0.72–0.86]	0.87a [0.72–1]	n.d.	n.d.
MoCA score	n=12	n=12	n=12
MoCA total score	22.7b (5.1)	24.6b (4.4)	25.8b (4.2)	1.16 [0.05 to 2.26]	0.04
MoCA executive function	3.0b (1.5)	3.0b (1.4)	3.4b (1.7)	0.36 [−0.02 to 0.74]	0.06
Grip strength
Grip strength, kg	25.5b [17.4–34.8]	25b [16.5–33.5]	21.5b [18–34]	−0.57 [−4.14 to 2.99]	0.68
Gait tests
Gait speed normal walk, cm/s	99.7b (35.5)	100.6b (32.3)	106.0c (25.9)	−0.8 [−5.9 to 4.3]	0.77
Step width normal walk, cm	9.9b (4.2)	10.9b (4.1)	10.0c (5.0)	−0.8 [−2.2 to 0.5]	0.28
Cycle time variability normal walk, %	3.0b (1.3)	3.0b (1.6)	2.3c (0.6)	−0.21 [−0.78 to 0.35]	0.48
Single support time normal walk, s	0.45b (0.11)	0.44b (0.06)	0.41c (0.03)	−0.0062 [−0.0138 to 0.0015]	0.16

These data are from 14 patients (unless otherwise indicated). Summary statistics are shown as median [interquartile range] or mean (standard deviation) for descriptive purposes. For inferential purposes, the treatment effect is indicated by the estimated difference in the outcome (estimate [95% CI]: empagliflozin–placebo), derived from linear mixed-effects models fit by maximum likelihood. Models included the baseline levels as covariate and patient as random effect. Each outcome was analyzed by a separate statistical model. The treatment effect is calculated from the individual, within-patient differences empagliflozin–placebo and is adjusted for baseline levels; the treatment effect cannot be derived directly from the summary statistics. S, serum; U, urinary; VAS, visual analog scale; n.d., not done.

^a13 patients.

^b12 patients.

^c10 patients.

Efficacy

After 4 weeks of treatment with empagliflozin, serum sodium levels rose to a median of 134 mmol/L (IQR, 132–136), whereas no improvement was seen under placebo (130 mmol/L; IQR, 127.5–132) (Table 2). This resulted in an estimated increase of 4.09 mmol/L (95% CI, 1.68 to 6.49) under empagliflozin as compared with placebo, (ITTS, P=0.004; Figure 2). This finding was confirmed by the PPS: 4.44 mmol/L (95% CI, 1.87 to 7.00; P=0.004). While serum sodium levels rose already within the first week of treatment with empagliflozin and stayed constantly increased for the remaining treatment period, no notable change was seen under placebo (Figure 2); the difference in AUC weeks 1–4 was 9.08 mmol/L per 3 weeks (95% CI, 2.78 to 15.39; P=0.01).

Figure 2.

Course of serum sodium levels from baseline to end of treatment. (A) Serum sodium levels before (baseline) and after (week 4) treatment with empagliflozin or placebo. (B) Course of serum sodium from baseline to end of treatment according to treatment phase. Boxes contain the 25% and 75% quartiles (spanning the interquartile range); the thick horizontal line is the median. Whiskers indicate the most extreme values lying within the box edge and 1.5 times the interquartile range. All eventual further values are plotted as individual points (outliers).

Under treatment with empagliflozin, 36% (5 of 14) of the patients reached normonatremia compared with 14% (2 of 14) of the patients under placebo. One patient reached normonatremia under both treatments. Six patients had moderate hyponatremia (serum sodium <130 mmol/L) at the start of treatment with empagliflozin, with two of these patients also having moderate hyponatremia at the start of placebo treatment. At the end of the treatment period, no patients on empagliflozin had moderate hyponatremia compared with six patients on placebo. Empagliflozin was effective in increasing serum sodium levels independently of the randomization sequence, there was no carryover effect (Supplemental Figure 1). At the 30-day follow-up visit, 79% (11 of 14) of the patients showed persistent hyponatremia.

Fluid intake remained stable during the active study phase with no relevant difference between the treatment cycles (median [IQR]: fluid intake at week 1 was 1.5 L [1.3–1.6] for empagliflozin versus 1.6 L [1.4–1.9] for placebo; and at week 4 was 1.5 L [1.2–1.5] for empagliflozin versus 1.5 L [1.1–1.5] for placebo).

Patients with a higher fractional excretion of urea at baseline tended to have a higher increase in serum sodium at the end of the treatment period, irrespective of the treatment (estimate: 0.122 mmol/L; 95% CI, 0.007 to 0.236; P=0.059). No evidence of an association was found for outcome or treatment effect with serum or urine osmolality, or urine sodium, or fractional excretion of sodium, or uric acid (Supplemental Figure 2) and for age (data not shown).

Secondary Outcomes

The course of clinical and laboratory parameters and the different assessments are summarized in Table 2.

Clinical and Laboratory Parameters

As expected, a strong increase in urine glucose and osmolality was observed after 4 weeks of treatment with empagliflozin compared with placebo (Figure 3 and Supplemental Figure 3). Additionally, treatment with empagliflozin compared with placebo led to an increase in serum osmolality and mild decrease in serum glucose and body weight. Our data further showed a slight increase in serum creatinine under treatment with empagliflozin. Changes during the treatment phase occurred within the first week and were maintained until the end of treatment (Figure 3). No effect of empagliflozin was seen for the other clinical or laboratory parameters.

Figure 3.

Time course of different parameters under treatment with empagliflozin or placebo. Results are shown for (A) serum glucose, (B) serum creatinine, (C) urine glucose, and (D) urine sodium. Boxes contain the 25% and 75% quartiles (spanning the interquartile range); the thick horizontal line is the median. Whiskers indicate the most extreme values lying within the box edge and 1.5 times the interquartile range. All eventual further values are plotted as individual points (outliers).

Assessments

The patients rated their general health state as rather good according to the EQ-5D-5L questionnaire. Compared with baseline, the majority of patients indicated an improved overall health after both treatment cycles, with no relevant difference between treatments.

The mean (SD) score in the MoCA at baseline was 22.7 (5.1). After 4 weeks of treatment with empagliflozin, the mean (SD) score rose to 25.8 (4.2) compared with 24.6 (4.4) under placebo. This resulted in a difference of 1.16 (95% CI, 0.05 to 2.26) under empagliflozin compared with placebo. A similar observation was made for the MoCA executive function subscore with a difference of 0.36 (95% CI, −0.02 to 0.74). Visualization and evaluation of the data showed no signs of a possible learning effect (P=0.29) or sequence effect (P=0.21), (Supplemental Figure 4).

Test scores after both treatment phases of patients reaching normonatremia (seven observations) were compared with those of patients with persistent hyponatremia (17 observations). Although no effect of sodium normalization on MoCA total score was seen (0.29; 95% CI, −1.45 to 2.02), a beneficial effect was seen for the MoCA executive function subscore (0.77; 95% CI, 0.16 to 1.38).

No notable change from baseline or difference between empagliflozin and placebo was seen in grip strength (Table 2), independently of whether treatment led to normonatremia or not (data not shown).

Baseline gait analysis showed slightly slowed walking speed with a healthy step width, single support time, and regular gait, as measured by the cycle time variability. Although some improvements were observed under empagliflozin resulting in faster gait, reduced cycle time variability, and shorter single support time, no difference was seen between treatments. Also, no effect of sodium normalization could be shown.

Tolerability and Safety

Treatment with empagliflozin was generally well tolerated (Table 3). Of the evaluated symptoms, thirst was indicated in about half of all patients under both treatments and at any measurement time. Few patients reported headache and vertigo; nausea was only reported under placebo.

Table 3.

Symptoms and adverse events occurring during observation phase

Characteristic	Placebo		Empagliflozin
	Week 0	Week 4	Week 0	Week 4
Symptoms: yes, n (%)
Thirst	8 (57)	6 (43)	6 (43)	9 (64)
Vertigo	1 (7)	1 (7)	1 (7)	3 (21)
Headache	1 (7)	2 (14)	5 (39)	3 (21)
Nausea	0 (0)	2 (15)	0 (0)	0 (0)
Adverse events, n
All adverse events	7		7
Serious adverse events	0		0
Potentially study related	2		5
Specific adverse events, n
Mild headache	1		1
Potentially study related	1		1
Viral/bacterial/fungal infection	4		1
Potentially study related	0		0
Gastrointestinal disorders	0		2
Potentially study related	0		2
Dry mouth	0		1
Potentially study related	0		1
Hypertensive episode	1		0
Potentially study related	0		0
Tiredness	0		1
Potentially study related	0		1
Gait insecurity (heavy legs)	0		1
Potentially study related	0		0
Exanthema	1		0
Potentially study related	1		0

n, number of patients.

No events of sodium overcorrection, hypoglycemia, hypotension, or urinary tract or genital infection occurred during the observation period under empagliflozin. Of the seven reported adverse events, five were potentially related to empagliflozin treatment. Under placebo, two of the reported seven adverse events were judged to be potentially related to the study intervention. No adverse events were recorded for the two participants who withdrew their consent. There were no serious adverse events during the observation period.

Discussion

We here show that treatment with the SGLT2 inhibitor empagliflozin leads to a relevant increase in serum sodium levels compared with placebo in outpatients with chronic SIAD. Exploratory analyses suggest a possible treatment-induced improvement in neurocognitive function. Treatment was safe and well tolerated.

In addition to the findings of our proof-of-concept study consisting of a 4-day treatment period in hospitalized patients with SIAD,¹⁸ we here show that empagliflozin is efficient as a long-term treatment option for chronic SIAD-induced hyponatremia. The rise in serum sodium levels was similar to the one of two recent intervention studies, evaluating the efficacy of fluid restriction of <0.5–1 L/d.^8,31 However, such a pronounced fluid restriction is a burdensome treatment that is unlikely to be sustained over an extended period. Meanwhile, the effect of empagliflozin was evident within one week of treatment with sodium levels remaining consistently elevated until the end of the treatment cycle, supporting the efficacy of a long-term treatment effect. Possibly, the treatment effect of empagliflozin could be increased by a mild fluid restriction analogous to our previous study.¹⁸

Having a higher fractional excretion of urea was found to be a predictive marker for empagliflozin treatment response in our study. Because this ratio is usually used to distinguish between SIAD and hypovolemic hyponatremia,³² a higher value likely characterizes patients with more substantial water excess that therefore benefit from osmotic diuresis. Interestingly, all patients for which serum sodium levels did not increase under placebo (change ≤0) showed a clear increase in serum sodium levels after empagliflozin treatment. This observation further strengthens the role of empagliflozin as a new treatment option in patients with chronic SIAD.

One safety concern of the proof-of-concept study¹⁸ was the transient decrease in renal function observed in four patients in the empagliflozin group. A mild 6% decrease in glomerular function was again observed in this study. However, several large outcome studies investigating SGLT2 inhibitors in diabetic and nondiabetic patients^23,33,34 showed that a decrease in GFR of up to 30% of baseline after initiation of treatment is to be expected and was associated with nephroprotective effects. Other factors associated with improved outcome were glucosuria and mild weight loss, which were also present in our cohort. Accordingly, we consider treatment with empagliflozin to be safe in patients with chronic SIAD-induced hyponatremia, as also reflected by the low number of only mild adverse events.

In addition to being a novel, well-tolerated, and efficient treatment option for chronic SIAD-induced hyponatremia, exploratory analyses also suggest a possible treatment-induced improvement in neurocognitive function. Several studies report neurocognitive deficits in hyponatremic patients, regardless of the severity of hyponatremia.^15–17 This was confirmed in our cohort with mild to moderate hyponatremia, where patients achieved a median MoCA score of 22.7 (normal ≥26 points) indicating mild to moderate impairment. However, despite one tolvaptan trial showing improvement in the psychomotor speed domain (a subtest combining neurocognitive and gait function⁷) no interventional study has yet demonstrated reversibility of neurocognitive impairments after treatment of hyponatremia.^35–38 The MoCA test is very sensitive for mild cognitive impairment, which could explain why we were able to detect an effect. We consider the observed improvement to be significant, because most intervention studies (even in other disorders outside the area of hyponatremia) showed no effect on MoCA scores or only in patients with severe neurocognitive impairment.^39–42 One study evaluating physical exercise was able to show a two-point increase after 3 months of intensive training.⁴³ Normonatremic patients in our study reached a higher score on the executive function subtest of the MoCA. This suggests the increase in sodium levels as an underlying mechanism for improvement of neurocognitive function. Although a drug-specific effect of empagliflozin cannot be excluded, it seems unlikely in view of several observational studies showing improvement in neurocognitive function after treatment, especially in those patients reaching normonatremia.^15,17,44

Chronic hyponatremia has been associated with gait instability leading to increased falls^17,45,46 and two observational studies have shown gait normalization after hyponatremia correction.^17,45 The baseline gait parameters of this study population showed only mild gait disturbances, primarily manifested with reduced walking speed. Although some improvements in gait were noted after treatment, no differences were seen between the interventions. The fact that no clinically relevant improvement was observed in our cohort could be due to the good baseline situation. Another explanation could be that, compared with the two observational studies, only a minority of our patients achieved normonatremia. Interestingly, however, a post- versus pretreatment reduction in gait cycle time variability was observed, suggesting improved gait regularity. Because deficits in executive function are associated with increased gait cycle time variability,⁴⁷ this finding is in line with the improvement in the MoCA executive function subscore.

With daily treatment cost of empagliflozin being similar to urea (approximately 2 USD versus 4 USD), but being 1/40th of the daily cost of tolvaptan (approximately 80 USD), empagliflozin would also be a cost-effective treatment option. Considering the cardiovascular and renal benefits of SGLT2 inhibitors, a chronic treatment with empagliflozin can be seen as a holistic approach in patients with chronic hyponatremia, who are usually older and have a high burden of comorbidities.

The strength of our study lies in the prospective randomized, double-blind, crossover design and the novelty of the treatment approach. The observed treatment effect of empagliflozin is convincing and should prompt further evaluations of empagliflozin as an efficient treatment option in SIAD-induced hyponatremia.

A limitation of our study includes the small patient population and that not all patients were able to perform the planned assessments (due to comorbidities and the ongoing COVID-19 pandemic), thereby inducing a possible power issue for detecting a difference in test performance. Also, although our data suggest an increase in electrolyte free water excretion as an explanation for the therapeutic effect of empagliflozin on serum sodium levels, this cannot be verified because urine output and urine potassium were not measured.

In conclusion, this study shows that the SGLT2 inhibitor empagliflozin is a promising treatment option for outpatients with chronic SIAD-induced hyponatremia, possibly leading to an improvement of neurocognitive function. Larger studies in in- and outpatient settings are needed to confirm the observed treatment effects.

Supplementary Material

jasn-34-322-s001.pdf ^{(623.4KB, pdf)}

Disclosures

All authors have nothing to disclose.

Funding

This study was supported by the Swiss Endocrine Society (Young Investigator grant to J. Refardt); by the University Hospital Basel and Gottfried and Julia Bangerter-Rhyner Stiftung (Young Talents in Clinical Research grant to C. Imber); and by the Swiss National Science Foundation (SNF-162608 to M. Christ-Crain; SNF-199391 MD-PhD fellowship to S. Monnerat). The funders had no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication.

Author Contributions

M. Christ-Crain, J. Refardt, D.R. Vogt, and B. Winzeler conceptualized the study; M. Christ-Crain was responsible for funding acquisition and validation; S.A. Bridenbaugh and M. Christ-Crain were responsible for supervision; C. Bathelt, S.A. Bridenbaugh, A. Haslbauer, C. Imber, R. Nobbenhuis, J. Refardt, and C.O. Sailer were responsible for investigation; C. Bathelt, A. Haslbauer, C. Imber, R. Nobbenhuis, J. Refardt, and C.O. Sailer were responsible for data curation; C. Bathelt, M. Christ-Crain, and J. Refardt were responsible for project administration; M. Christ-Crain, J. Refardt, and B. Winzeler were responsible for methodology; M. Berres, S. Monnerat, J. Refardt, and D.R. Vogt were responsible for formal analysis; S. Monnerat was responsible for resources; J. Refardt wrote the original draft; and C. Bathelt, M. Berres, S.A. Bridenbaugh, M. Christ-Crain, A. Haslbauer, C. Imber, S. Monnerat, R. Nobbenhuis, C.O. Sailer, D.R. Vogt, and B. Winzeler reviewed and edited the manuscript.

Data Sharing Statement

The following data will be shared upon publication to researchers who provide a methodologically sound proposal to achieve the aims in the approved proposal: deidentified individual participant data that underlie the results reported in this article; study protocol; and statistical analysis plan. Proposals should be directed to the corresponding author. To gain access, data requestors will need to sign a data access agreement.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/D725.

Supplemental Statistical Analyses

Supplemental Figure 1. Individual serum sodium measurements at baseline and at the end of treatment for each patient under each treatment, grouped by the two randomization sequences.

Supplemental Figure 2. Time course of fractional excretion of sodium, urea, and uric acid under treatment with empagliflozin or placebo.

Supplemental Figure 3. Time course of serum osmolality and urine osmolality under treatment with empagliflozin or placebo.

Supplemental Figure 4. Visualization of the MoCA total scores according to the two treatment cycles.