Methods

• Design: parallel RCT (phase 2)

• Hyponatraemia subgroup: yes, but randomisation not stratified according to the subgroup

• Recruitment period: 2000 to 2001

• Treatment duration: 60 days

• Follow‐up: 60 days

Participants

• Countries: Argentina; USA

• Number of centres: 44

• Inclusion criteria: in‐patients, hospitalised for acute exacerbation of CHF; ≥ 18 years; serum sodium concentration was not an inclusion criterion for primary study ‐ we included subgroup with serum sodium concentration ≤ 135 mmol/L; NYHA class III or IV cardiac disease at screening ‐ 2 of the following signs of heart failure: elevated jugular vein distention; rales; chest X‐ray with signs of radiographic congestion, pedal oedema, increased abdominal girth, weight gain > 10 pounds above baseline; admission to hospital for heart failure within 96 hours of screening

• Exclusion criteria: women of childbearing age; cardiac surgery within 60 days; MI, sustained ventricular tachycardia, or ventricular fibrillation within 30 days; angina at rest; primary valvular disease; hypertrophic cardiomyopathy; stroke within last 6 months; significant hepatic, renal or haematologic dysfunction; systolic arterial BP < 110 mm Hg; use of drugs known to inhibit CYP3A4 within 7 days of randomisation; amiodarone within 10 weeks; nonsteroidal anti‐inflammatory agents or aspirin > 700 mg/d; substance or alcohol abuse; uncontrolled diabetes mellitus; urinary tract obstruction; morbid obesity; malignancy or another terminal illness.

• Characteristics randomised participants

  • Number: treatment group 1 (15); treatment group 2 (23); treatment group 3 (15); control group (16)

  • Number, % women: 23, 33%; numbers not reported for individual groups

  • Age: median 65, range 52‐72 years; age not reported for individual groups

  • Mean serum sodium concentration ± SD (mmol/L): treatment group 1 (133 ± 3); treatment group 2 (131 ± 4); treatment group 3 (132 ± 3); control group (133 ± 2)

  • Cause of hyponatraemia: heart failure (100%)

Interventions

Treatment group 1

• Oral tolvaptan: 1 x 90 mg/d

Treatment group 2

• Oral tolvaptan: 1 x 60 mg/d

Treatment group 3

• Oral tolvaptan: 1 x 30 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: no

• Dietary sodium intake: no

• All received standard heart failure treatment including diuretics at the discretion of the investigator

Outcomes

• Death

• Response in serum sodium concentration: serum sodium concentration > 135 mmol/L. Insufficiently reported to allow contribution to meta‐analysis

• Serum sodium concentration (continuous outcome): absolute values at discharge

Notes

• Funding source: Otsuka Maryland Research Institute

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'Eligible patients were randomised ... using an interactive voice recognition system programmed with a computer‐generated randomisation scheme. Randomization was stratified by study centre in blocks of 4.'

Comment: the randomisation procedure produced groups in the overall study that were not entirely balanced in age and sex, but fairly well‐balanced in terms of medical history and characteristics related to the heart failure. However, randomisation was not stratified for serum sodium concentration. Despite fairly similar numbers of participants in both the experimental and the control group, no baseline data available specifically for subgroup with hyponatraemia, making judgement on risk of bias difficult. The sequence generation in the overall study was good and low risk. However this is a subgroup analysis with low numbers relative to the number of groups and the sequence did not stratify for them.

Allocation concealment (selection bias)

Low risk

Quote: 'Patients were screened within 72 hours and randomised within 96 hours of admission. Randomization occurred between 8 and 9 AM; the study drug was to be administered at 9 AM.'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: '...patients...randomised to receive ... tolvaptan or placebo. ' and '...was a multicenter, randomised, double‐blind, placebo‐controlled... trial'

Additional info provided by the sponsor: 'Both OPC‐41061 and placebo tablets were identical in appearance'.

Comment: blinding attempted, although participants are unlikely to be fully blinded due to important increases in urine output. Also, co‐interventions were not reported in detail and included standard heart failure treatment left at discretion of investigator and may have differed between groups according to baseline serum sodium concentration. This could have influenced outcome measurement of serum sodium concentrations.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Additional info provided by the sponsor: 'Did not have separate outcome assessors in this trial'.

Comment: blinding outcome assessors attempted and measured outcome objective.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: 10% attrition for outcome death at 60 days; only 1 participant (1%) excluded from analysis because of missing follow‐up serum sodium concentration measurement.

Selective reporting (reporting bias)

High risk

Comment: protocol available with matched primary and secondary outcomes for the overall study. However, results included in this review for the subgroup with hyponatraemia are based on post‐hoc analyses of the data, with reported outcomes not matching the primary and secondary outcomes of the original study.

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • 'All authors served as consultants for Otsuka Maryland Research Institute, received grants or honoraria from the company, or both.'

  • Comment: senior authors of reports pertaining to this study are heavily linked to the company commercialising the study medication.

• Sponsorship bias

  • Quote: 'The ACTIV in CHF study and this analysis were funded by Otsuka Maryland Research Institute'

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 5 days

• Follow‐up: 12 days

Participants

• Countries: Belgium; Finland; France; Germany; Italy; Netherlands; Poland; Spain; UK

• Number of centres: 17

• Inclusion criteria: aged ≥ 18 years; serum sodium concentration < 130 mmol/L; euvolaemia or hypervolaemia, plasma osmolality < 290 mOsmol/kg, fasting blood glucose < 126 mg/dL

• Exclusion criteria: females if breast‐feeding or pregnant; clinical volume depletion or dehydration; symptomatic hypotension, uncontrolled hypertension, arrhythmias requiring pacemaker or emergency treatment; untreated thyroid disorders, adrenal insufficiency; known urinary outflow obstruction unless catheterized; SCr >2.7 mg/dL, proteinuria > 3 g/24h, liver enzymes > 3 times upper limit of normal; bilirubin ≥ 2.5 mg/dL, < 3x 10³ white blood cells/µL, humane immuno‐deficiency virus infection or active hepatitis; symptomatic hyponatraemia, expected to require emergent treatment; any concurrent illness that could interfere with study treatment or its evaluation; arginine vasopressin, oxytocin, or desmopressin or drugs interacting with CYP450 3A4, including certain chemotherapeutic agents, calcium channel blockers, statins, benzodiazepines, and class III antiarrhythmic drugs; medications for SIADH discontinued for 1 week before screening; thiazide diuretics permitted if therapy initiated < 1 month before study enrolment

• Characteristics randomised participants

  • Number: treatment group 1 (26); treatment group 2 (27); control group (30)

  • Number, % women: treatment group 1 (7, 27%); treatment group 2 (9, 33%); control group (12, 40%)

  • Mean age ± SD (years): treatment group 1 (63 ± 11); treatment group 2 (62 ± 15); control group (69 ± 13)

  • Mean serum sodium concentration ± SD (mmol/L): treatment group 1 (126 ± 4); treatment group 2 (125 ± 5); control group (126 ± 4)

  • Cause of hyponatraemia

    • Treatment group 1: SIADH (12, 46%); heart failure (7, 27%); unclear (7, 27%)

    • Treatment group 2: SIADH (7, 26%); heart failure (12, 44%); unclear (8, 30%)

    • Control group: SIADH (12, 40%); heart failure (8, 40%); unclear (7, 23%); postsurgical (3, 10%)

Interventions

treatment group 1

• Oral conivaptan: 2 x 40 mg/d

Treatment group 2

• Oral conivaptan: 2 x 20 mg/d

Control group

• Placebo: not otherwise defined

Co‐interventions (all groups)

• fluid restriction: 500 mL in any 3‐hour period or 2 L/d, unless changes required for safety reasons

• dietary sodium intake: not reported

Outcomes

• Death

• Response: serum sodium concentration ≥ 135 mmol/L or increase from baseline ≥ 6 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline until end of treatment

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: serum sodium concentration ≥ 145 mmol/L during treatment

  • Rapid increase in serum sodium concentration: serum sodium concentration ≥ 145 mmol/L or an increase in serum sodium concentration from baseline ≥ 12 mmol/L/d

• Treatment‐specific side‐effects

  • Hypotension: not otherwise defined

  • Thirst: not otherwise defined (no extractable data reported)

  • Liver function abnormalities: elevated liver transaminase

  • Serious adverse event: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Astellas Pharma US, Inc.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'Eligible patient ... were stratified according to volume status and then randomly assigned in a 1:1:1 ratio...'

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'This double‐blind, placebo controlled ... study', and '...patients ... were...assigned to ... or placebo', and 'All patients were prescribed fluid restriction (not to exceed 500 mL in any 3‐hour period or 2.0 L in a 24‐hour period for both oral and parenteral fluids) prior to baseline and throughout the study, unless changes were necessary for safety reasons', and 'Four patients, 2 given conivaptan 80 mg/d and 2 given placebo, violated the 2‐L/d fluid restriction.'

Comment: blinding attempted, and although participants and personnel are unlikely to be fully blinded due to important increases in urine output, all measured outcomes short‐term and generally objective. Fluid restriction was prescribed in both groups, violations reported and similar between groups

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding outcome assessors probably attempted and main outcomes objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 13% attrition; reasons well documented, most occurred in the placebo group due to adverse events. Given <20% attrition and limited opportunity for introducing bias in such a short‐term study ‐ we judged it 'Low risk'.

Selective reporting (reporting bias)

Low risk

Comment: no protocol, short term study (≤1 week), primary outcome death and secondary outcomes related to serum sodium concentration and rapid increase in serum sodium concentration reported

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'Guy Decaux was a paid investigator for this trial. Neila Smith was employed by Astellas Pharma US, Inc.

  • Comment: Senior author was an employee of the company commercialising the study medication

• Sponsorship bias

  • 'This study was supported by Astellas Pharma US, Inc.'

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: 2008 to 2010

• Treatment duration: 60 days

• Follow‐up: 90 days

Participants

• Country: Argentina; Canada; Chile; Czech Republic; Germany; India; Israel; Italy; Poland; Romania; Russia; Slovakia; Spain; USA

• Number of centres: 173

• Inclusion criteria: inpatients, hospitalised for acute worsening of CHF; ≥ 18 years; serum sodium concentration 120 to < 135 mmol/L; NYHA III or IV; clinical or laboratory evidence of volume overload, baseline fluid restriction 1.5 L/d

• Exclusion criteria: severe symptoms requiring immediate intervention; supine systolic BP < 90 mm Hg; advanced liver disease/cirrhosis; adrenal insufficiency; uncorrected thyroid disease; uncontrolled diabetes mellitus; inability to respond to thirst with increased fluid intake

• Characteristics randomised participants

  • Number: treatment group (323); control group (329)

  • Number, % women: treatment group (90, 28%); control group (95, 29%)

  • Mean age ± SD (years): treatment group: (65 ± 14); control group (65 ± 13)

  • Serum sodium concentration: median 132 mmol/L; concentration not reported for groups

  • Cause of hyponatraemia: heart failure (100%)

Interventions

Treatment group

• Oral lixivaptan: 1 x 50 mg/d up to 2 x 100 mg/d

Control group

• Oral placebo: 1 x 1 capsule/d

Co‐interventions (all groups)

• Fluid restriction: discouraged during dose titration, at discretion of treating physician

• Dietary sodium intake: not reported

Outcomes

• Death

• Response: serum sodium concentration ≥ 135 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline

• Cognitive function: change from baseline time to complete the trial making test part B, a neuropsychological test of visual attention and task switching

• Outcomes related to over‐correction of serum sodium concentration

  • Rapid increase in serum sodium concentration: serum sodium concentration increase ≥ 8 mmol/L during first 8 hours, > 12 mmol/L during first 24 hours or > 18 mmol/L during first 48 hours

• Treatment‐specific side‐effects

  • AKI: not otherwise defined

  • Liver function abnormalities: elevated AST > 3 times the upper limit of normal

  • Serious adverse events: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Cardiokine Inc.

• ClinicalTrials.gov: NCT00578695

• Data source: FDA Briefing Document for the Cardiovascular and Renal Drugs Advisory (CRDAC)

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: '...subjects were randomised by an Interactive Voice Response System...'.

Comment: central allocation occurred.

Allocation concealment (selection bias)

Low risk

Quote: '...subjects were randomised by an Interactive Voice Response System...'.

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Quote: '... double‐blind, placebo‐controlled phase 3 trial...'

Comment: Blinding of participants and personnel probably attempted, but insufficient information to assess the extent to which it was achieved or to permit judgement of 'High risk' or 'Low risk'.. Co‐interventions were left at the discretion of the treating physician.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

'The Clinical Endpoints Committee was to perform a review (blinded to treatment group and serum sodium levels) and adjudicate all events assessed by the investigators as clinical endpoints using predetermined criteria for each end‐point...'.

Comment: blinding probably attempted and although un‐blinding likely to have occurred for serum sodium concentration, assessment of other outcomes done by independent blinded adjudication committee.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Quote: 'The FDA statistical reviewer, Dr Jialu Zhang, also performed a number of sensitivity analyses to assess the impact of missing values... All these sensitivity analyses showed consistent results: a statistically significant treatment effect on serum sodium with a modest effect size in the range 1.2 to 1.4 mEq/L.'

Comment: overall 27% discontinuation at 30 days, < 1% of which never received treatment; 50% of discontinuations occurred in placebo group. Last observation carried forward method was used as imputation method for dealing with missing serum sodium concentration values, which could have caused overestimation of the effect of the study medication. However, sensitivity analyses conducted by the FDA statistical reviewer indicated this effect to be negligible. Cognitive function tests were taken at day 30 or the early termination visit. Given the better scores in the placebo group, any bias would only serve to increase the difference further in favour of the placebo group and would not change the conclusions drawn from the analysis.

Selective reporting (reporting bias)

Low risk

Comment: the study was registered with ClinicalTrials.gov, but no clear listing of predefined outcome measures and time‐points. Given early stopping of trial due to numerically more deaths in treatment group, selective outcome reporting unlikely to be an issue.

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • No declaration of interest: data used in this review is drawn from the FDA briefing document

• Sponsorship bias

  • Comment: The study sponsor was Cardiokine Inc.

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 4 days

• Follow‐up: 1 year

Participants

• Countries: not reported

• Number of centres: not reported, but multi‐centre

• Inclusion criteria: setting (not reported); age (not reported); serum sodium concentration 115 to ≤ 132 mmol/L; SIADH treated with fluid restriction up to 1.5 L/d

• Exclusion criteria: not reported

• Characteristics included participants

  • Number: treatment group 1 (22); treatment group 2 (24); control group (23)

  • Number, % women: not reported

  • Age: not reported

  • Mean serum sodium concentration ± SD (mmol/L): treatment group 1 (129 ± 3); treatment group 2 (129 ± 5); control group (128 ± 4)

  • Cause of hyponatraemia: SIADH not reported

Interventions

Treatment group 1

• Oral satavaptan: 1 x 50 mg/d

Treatment group 2

• Oral satavaptan: 1 x 25 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: not reported, participants were treated with fluid restriction at enrolment

• Dietary sodium intake: not reported

Outcomes

• Response: serum sodium concentration ≥135 mmol/L or increase from baseline ≥ 5 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline. Insufficiently reported to allow contribution to meta‐analysis

• Outcomes related to over‐correction of serum sodium concentration

  • Rapid increase in serum sodium concentration: serum sodium concentration increase ≥ 8 mmol/L during the first day

  • Osmotic demyelination syndrome: any neurological adverse event

• Treatment‐specific side‐effects: insufficiently reported to allow contribution to meta‐analysis

  • Thirst: not otherwise defined

  • Adverse events: defined as treatment‐emergent adverse events

  • Serious adverse events: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Sanofi Aventis

• Publication: abstract only; contact author confirmed the study was never published in any form other than abstract form

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'randomised... study', and 'Patients ... were randomly assigned'.

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Quote: '...double‐blind, placebo‐controlled...study', and 'Patients...were... assigned to ... or placebo'.

Comment: blinding probably attempted, but insufficient information to assess to what extent it was achieved or to permit judgement of 'High risk' or 'Low risk'.

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: overall 16% attrition, reasons not documented. Given < 20% attrition in short‐term study, we judged this not at high risk of bias per se, rather insufficient information to permit judgement of 'High risk' or 'Low risk'.

Selective reporting (reporting bias)

Low risk

Comment: no protocol, short‐term study (≤ 1 week), secondary outcomes related to serum sodium concentration and over‐correction of serum sodium concentration reported.

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • No declaration of interest

• Sponsorship bias

  • Quote: 'Grant/Research Support: Sanofi‐Aventis; Consultant: Sanofi‐Aventis

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 4 days

• Follow‐up: not reported

Participants

• Countries: Argentina; Australia; Belgium; Canada; Chile; Denmark; Greece; Hungary; Israel; Poland; Portugal; Romania; South Africa; Sweden; USA

• Number of centres: 48

• Inclusion criteria: setting not reported; ≥ 18 years; serum sodium concentration 115 to < 132 mmol/L; dilutional hyponatraemia not caused by SIADH or liver cirrhosis

• Exclusion criteria: adrenal insufficiency; untreated hypothyroidism; clinical or laboratory evidence of hypovolaemia; fasting glycaemia ≥ 200 mg/dL; Hb < 9 g/dL, neutrophil cell count < 1500/µL, platelet count < 100,000/µL; SCr > 2 mg/dL or CrCl < 30 mL/min; serum ALT or AST > 2 times upper limit of normal; QTcB interval ≥ 500 ms; serum potassium concentration > 5 mmol/L; having received inducers of CYP3A4 or potent and moderate inhibitors of CYP3A4 < 2 weeks before study; other AVP V2‐receptor antagonists, lithium or demeclocycline < 1 month before study, urea < 2 days before study; positive pregnancy test and absence of medically approved contraceptive methods; pregnancy or breast‐feeding

• Characteristics randomised participants

  • Number: treatment group 1 (41); treatment group 2 (35); control group (42)

  • Number, % women: treatment group 1 (16, 39%); treatment group 2 (19, 54%); control group (15, 36%)

  • Mean age ± SD (years): treatment group 1 (71 ± 12); treatment group 2 (69 ± 16); control group (69 ± 11)

  • Serum sodium concentration (mmol/L): treatment group 1 (128 ± 4); treatment group 2 (128 ± 4); control group (129 ± 3)

  • Cause of hyponatraemia

    • Treatment group 1: heart failure 28 (68%); other 13 (32%)

    • Treatment group 2: heart failure 28 (80%); other 7 (20%)

    • Control group: heart failure 34 (81%); other 8 (19%)

Interventions

Treatment group 1

• Oral satavaptan: 1 x 50 mg/d

Treatment group 2

• Oral satavaptan: 1 x 25 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: 1 to 1.5 L/d

• Dietary sodium intake: not reported

Outcomes

• Response: serum sodium concentration ≥ 135 mmol/L or increase from baseline ≥ 5 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: serum sodium concentration ≥ 145 mmol/L

  • Rapid increase in serum sodium concentration: serum sodium concentration increase ≥ 12 mmol/L/d

  • Osmotic demyelination syndrome: neurologic symptoms

• Treatment‐specific side‐effects

  • Hypotension: not otherwise defined

  • Thirst: not otherwise defined

  • Adverse event: not otherwise defined

  • Serious adverse events: defined as serious treatment‐emergent adverse events

• Treatment discontinuation

Notes

• Funding source: Sanofi‐Aventis

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote:'...eligible patients were randomly allocated to...', and '4 patients were not randomised through the interactive voice recognition system'.

Comment: Central allocation probably occurred. The randomisation produced groups well‐balanced in age and serum sodium concentration, less so in sex

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: '...patients were randomly allocated to ... or placebo', and 'The randomised, double‐blind, placebo‐controlled ...', and '...recommended fluid intake 1000‐1500 mL'.

Comment: blinding attempted, and although participants and personnel are unlikely to be fully blinded due to important increases in urine output, all measured outcomes short‐term and fairly objective. Fluid restriction was prescribed in all groups

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding outcome assessors probably attempted and all measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 11% attrition, due to treatment discontinuation, reasons fairly well documented, asymmetrically in experimental group receiving highest dose due to adverse events. Missing data dealt with using last observation carried forward method. Given < 20% attrition and limited opportunity for introducing bias in such a short‐term study ‐ we judged it 'Low risk'

Selective reporting (reporting bias)

Low risk

Comment: study registered with ClinicalTrials.gov in 2005. Primary outcome matches, more outcomes reported than originally registered ‐ predominantly outcomes related to harms. Short‐term study (≤ 1 week), secondary outcomes related to serum sodium concentration and rapid increase in serum sodium concentration reported

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • M.M. was an employee of Sanofi‐Aventis at the time of the study. All other listed authors have received research funds to conduct the DILIPO study.'

  • Comment: One of the authors was an employee of the company developing the study medication

• Sponsorship bias

  • Quote: 'This study was supported by a research grant from Sanofi‐Aventis

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 5 days

• Follow‐up: 8 days

Participants

• Country: USA

• Number of centres: 1

• Inclusion criteria: serum sodium concentration ≤135 mmol/L; heart failure ‐ NYHA class III‐IV

• Exclusion criteria: not reported

• Characteristics randomised participants

  • Number: treatment group (8); control group (6)

  • Number, % women: treatment group (2, 25%); control group (0, 0%)

  • Mean age, range: 58, 3 to 72 years

  • Mean serum sodium concentration ± SD (mmol/L): treatment group (129 ± 1); control group (130 ± 1)

  • Cause of hyponatraemia: heart failure (100%)

Interventions

Treatment group

• Oral captopril 3 x 12.5 mg/d up to 3 x 150 mg/d

• Furosemide: at doses adjusted to give a sustained diuretic effect without increasing azotaemia

Control group

• Oral captopril: 3 x 12.5 mg/d up to 3 x 150 mg/d

Co‐interventions (all groups)

• Fluid restriction: free access to water

• Dietary sodium intake: 0.9 g/d

Outcomes

• Serum sodium concentration (continuous outcome): absolute value at end of treatment

Notes

• Funding source: National Institutes of Health and National Aeronautics and Space Administration

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'The patients were randomly divided in two groups...'

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Comment: blinding not attempted

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding not attempted, but main outcomes objective such that un‐blinding of outcome detection unlikely to be an important source of bias

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: attrition or treatment discontinuation not specifically reported, but it appears as if all participants completed the treatment protocol

Selective reporting (reporting bias)

High risk

Comment: no protocol, short‐term study (≤ 1 week), secondary outcomes related to rapid increase in serum sodium concentration not reported

Other bias

Low risk

• Bias through possible financial conflict of interest of the authors

  • Comment: no declaration of interest

• Sponsorship bias

  • Quote: 'Grant support: supported in part by grants ... from the National Institutes of Health and ....from the National Aeronautics and Space Administration.'

  • Comment: no commercial sponsorship

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: yes, randomisation not stratified based on serum sodium concentration

• Recruitment period: 2003 to 2006

• Treatment duration: minimum 2 months

• Follow‐up: 2 years

Participants

• Countries: Argentina; Belgium; Brazil; Bulgaria; Canada; Czech Republic; France; Germany; Italy; Lithuania; Netherlands; Norway; Poland; Romania; Russian Federation; Spain; Sweden; Switzerland; UK; USA

• Number of centres: 359

• Inclusion criteria: hospital in‐patients recruited within 48 hours of hospitalisation; ≥ 18 years; serum sodium concentration not an inclusion criterion for primary study (we included subgroup with serum sodium concentration < 135 mmol/L); history of CHF ‐ defined as requiring treatment for a minimum of 30 d before hospitalisation; left ventricular ejection fraction ≤ 40%; signs of volume expansion; NYHA III/IV; hospitalisation for exacerbation of chronic HF ≤ 48 h earlier

• Exclusion criteria: cardiac surgery < 60 days of enrolment; cardiac mechanical support; biventricular pacemaker < 60 days; comorbid conditions with expected survival < 6 months; acute MI at the time of hospitalisation; haemodynamically significant uncorrected primary cardiac valvular disease; refractory end‐stage heart failure; haemofiltration or dialysis; supine systolic arterial BP < 90 mm Hg; SCr > 3.5 mg/dL; serum potassium concentration > 5.5 mmol/L; Hb < 9 g/dL; history of hypersensitivity or idiosyncratic reaction to benzazepine derivatives; positive urine pregnancy test; inability to provide written informed consent; history of drug or medication abuse within the past year, or current alcohol abuse; previous participation in this or any other tolvaptan clinical trial; inability to take oral medications; participation in another clinical drug or device trial in which the last dose of drug was within the past 30 d or an investigation medical device is implanted

• Characteristics included participants

  • Number: treatment group (243); control group (232)

  • Number, % women: treatment group (35, 14%); control group (52, 22%)

  • Mean age ± SD (years): treatment group (64 ± 14); control group (65 ± 14)

  • Mean serum sodium concentration ± SD (mmol/L): treatment group (131 ± 3); control group (130 ± 4)

  • Cause of hyponatraemia: heart failure (100%)

Interventions

treatment group

• Oral tolvaptan: 1 x 30 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: not reported

• Dietary sodium intake: not reported

Outcomes

• Death during follow‐up

• Length of hospital stay

• Serum sodium concentration (continuous outcome): change from baseline until day 7

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: not otherwise defined

  • Osmotic demyelination syndrome: not otherwise defined

• Treatment‐specific side effects

  • AKI: not otherwise defined

  • Hypotension: not otherwise defined

  • Thirst (categorical outcome): not otherwise defined

  • Polyuria: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Otsuka

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: '...within each centre, patients were randomised to... according to a centralized, blocked randomization performed using a central Interactive Voice Response System'.

Additional info provided by the sponsor: 'Subjects were not stratified by baseline serum sodium'.

Comment: the randomisation procedure was adequate and produced well balanced groups in the overall study. Although randomisation was not stratified for serum sodium concentration, fairly similar numbers of participants in both the experimental and the control group, with similar ages and baseline serum sodium concentration

Allocation concealment (selection bias)

Low risk

Quote: '...within each centre, patients were randomised to... according to a centralized, blocked randomisation performed using a central Interactive Voice Response System'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: 'Patients...assigned to receive oral tolvaptan 30 mg or matching placebo. ' and '...is and international, ..., double‐blind, placebo‐controlled study...'.

Additional info provided by the sponsor: 'The trial drug was packaged so that each subject received an identical number of tablets regardless of the treatment group assigned'.

Comment: blinding attempted, although participants are unlikely to be fully blinded due to important increases in urine output. Also, co‐interventions were not reported and may have differed between groups according to baseline serum sodium concentration. This could have influenced outcome measurement of serum sodium concentrations and length of hospital stay

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Additional info provided by the sponsor: 'Did not have separate outcome assessors in this trial'.

Comment: blinding outcome assessors attempted and all measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

High risk

Comment: 30% (33% in experimental group, 29% in control group) attrition from analysis of serum sodium concentration at day 7

Selective reporting (reporting bias)

High risk

Comment: protocol available with matched primary and secondary outcomes for the overall study. However, results included in this review for the subgroup with hyponatraemia are based on post‐hoc analyses of the data, with reported outcomes not matching the primary and secondary outcomes of the original study

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'Ms Krasa received compensation through her Otsuka salary and Ms Bechhofer received compensation through a contract between the University of Wisconsin and Otsuka Marylad Research Institute.', and 'Mr. Cyr, Ms. Olchanski, Ms. Slawsky, and Dr. Gross received research funding from Otsuka America Pharmaceutical, which markets tolvaptan. Ms. Krasa and Dr. Zimmer are employees of Otsuka Pharmaceutical Development and Commercialization. Dr. Hauptman is a clinical investigator for and a consultant to Otsuka Pharmaceutical Development and Commercialization'

  • Comment: senior authors of reports pertaining to this study are heavily linked to the company commercialising the study medication

• Sponsorship bias

  • Quote: 'Otsuka funded the EVEREST trial under the guidance of the EVEREST steering committee', and 'the data collection and management for this study was by Otsuka; analysis was by the University of Wisconsin Statistical Data Analysis centre and Otsuka'.

  • Comment: Data collection, management and analysis was also done by the pharmaceutical company

Methods

• Design: parallel RCT (phase II)

• Hyponatraemia subgroup: no

• Recruitment period: 1997 to 1998

• Treatment duration: 7 days

• Follow‐up: not reported

Participants

• Country: not reported

• Number of centres: 17

• Inclusion criteria: 18 to ≤ 80 years; serum sodium concentration 115 to ≤ 132 mmol/L; urine/plasma osmolality > 1; written informed consent

• Exclusion criteria: major haemodynamic or pulmonary insufficiency; creatinine concentration >200 µmol/L; systolic arterial pressure < 80 mm Hg; Hb < 9 g/dL; encephalopathy; any evidence of active myocardial ischaemia within past two months; heart failure NYHA IV; recent antibiotic treatment

• Characteristics included participants

• Number: treatment group 1 (36); treatment group 2 (37); control group (39)

• Number, % women: not reported

• Mean age: 58 years (not reported for individual groups)

• Mean serum sodium concentration ± SD (mmol/L): treatment group 1 (128 ± 4); treatment group 2 (128 ± 4); control group (129 ± 4)

• Cause of hyponatraemia: SIADH (31, 28%); heart failure (14, 13%); liver cirrhosis (61, 54%) (not reported for individual groups)

Interventions

Treatment group 1

• Oral lixivaptan: 2 x 100 mg/d

Treatment group 2

• Oral lixivaptan: 2 x 50 mg/d

Control group

• Oral placebo: 2 x 1/d

Co‐interventions (all groups)

• Fluid restriction: < 1 L/d

• Dietary sodium intake: not reported

Outcomes

• Response: serum sodium concentration ≥ 136 mmol/L

• Serum sodium concentration (continuous outcome): absolute value at end of treatment

• Outcomes related to over‐correction of serum sodium concentration

  • Osmotic demyelination syndrome: any neurologic abnormality

• Treatment‐specific side‐effects

  • AKI: SCr > 200 µmol/L

  • Thirst (continuous outcome): measured on a 0 to 100 VAS

  • Serious adverse events: requiring treatment discontinuation

Notes

• Funding source: Wyeth Ayerst Research

• Publication: abstract‐only publication (full publication reports the subgroup of participants with cirrhosis only)

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: 'A computer‐generated randomisation table was available at each participating hospital. A patient number was assigned when a patient had qualified during the pre‐study screening. Before administration of the first study medication, the patient was assigned a randomisation number by an unblinded site pharmacist. This unblinded dispenser did not administer the drug to patients, and the contact between him and the investigators and study nurses was kept to a minimum.'

Comment: the randomisation procedure produced well balanced groups in terms of baseline serum sodium concentration, no information on other baseline characteristics

Allocation concealment (selection bias)

Low risk

Comment: the randomisation procedure produced well balanced groups in terms of baseline serum sodium concentration, no information on other baseline characteristics

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: '...randomised, double blind, placebo‐controlled study...', and '...patients were randomised to placebo or 50 mg twice a day...'

Comment: blinding attempted, and although participants and personnel are unlikely to be fully blinded due to important increases in urine output, all measured outcomes short‐term and fairly objective. Co‐intervention of fluid restriction similar in all groups

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: published in abstract form only, insufficient information to permit judgement of 'High risk' or 'Low risk'

Selective reporting (reporting bias)

High risk

Comment: no protocol, short‐term study (≤ 1 week) but no reporting of primary outcomes death or health‐related quality of life or secondary outcomes related to rapid increase in serum sodium concentration in a way that permitted data extraction

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'P Ginès has been a consultant for Wyeth Ayerst Research, H. Gandjini and J. Djian are employees at Wyeth Ayerst Research.'

  • Comment: senior author of all publications related to this study is an employee at the company developing the study medication

• Sponsorship bias

  • Comment: this study was funded Wyeth Ayerst Research

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 5 days

• Follow‐up: 14 days

Participants

• Countries: Canada; Israel; USA

• Number of centres: 21

• Inclusion criteria: ≥18 years; serum sodium concentration 115 to < 132 mmol/L; euvolaemia or hypervolaemia; plasma osmolality < 290 mOsm/kg

• Exclusion criteria: women if lactating, pregnant or not using barrier method of birth control; fasting serum glucose concentration ≥ 275 mg/dL; uncontrolled hypertension; orthostatic hypotension or supine systolic BP < 85 mm Hg; uncontrolled arrhythmia; untreated severe hypothyroidism, hyperthyroidism, adrenal insufficiency; estimated CrCl < 20 mL/min; ALT or AST > 5 times upper limit of normal; serum albumin concentration ≤ 1.5 g/dL; prothrombin time > 22 s or international normalised ratio > 2.0 without anticoagulants or ≥ 3.0 with therapy; WCC < 3000/µL; HIV infection or active hepatitis; patients expected to have hyponatraemia necessitating emergent treatment during study; concurrent illness that could interfere with study treatment or evaluation; participation in clinical trial with 30 days of screening

• • Characteristics included participants

  • Number: treatment group 1 (27); treatment group 2 (24); control group (23)

  • Number, % women: treatment group 1 (13, 48); %); treatment group 21 (2, 50%); control group (13, 56%)

  • Mean age, range (years): treatment group 1 (69, 35 to 90); treatment group 2 (66, 38 to 93); control group (73, 41 to 94)

  • Serum sodium concentration (mmol/L): treatment group 1 (125 ± 4); treatment group 2 (125 ± 4); control group (123 ± 4)

  • Cause of hyponatraemia

    • Treatment group 1: heart failure (11, 41%); COPD (2, 7%); malignancy (3, 11%); idiopathic (7, 26%); other (4, 15%)

    • Treatment group 2: heart failure (10, 42%); malignancy (2, 8%); idiopathic (5, 21%); other (7, 29%)

    • Control group: heart failure (11, 48%); COPD (1, 4%); malignancy (3, 13%); idiopathic (3, 13%); other (5, 22%)

Interventions

Treatment group 1

• Oral conivaptan: 2 x 40 mg/d

Treatment group 2

• Oral conivaptan: 2 x 20 mg/d

Control group

• Oral placebo: 2 x/d

Co‐interventions (all groups)

• Fluid restriction: ≤ 500 mL in any 3‐hour period or < 2L/d; treatment group 1 (67%), treatment group 2 (75%), control group (61%) drank 2 to 3 L/d

• Dietary sodium intake: remained stable throughout the study

Outcomes

• Death

• Response: serum sodium concentration ≥ 135 mmol/L or increase from baseline ≥ 6 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Outcomes related to over‐correction of serum sodium concentration

  • Rapid increase in serum sodium concentration: serum sodium concentration increase > 12 mmol/L/d or total increase > 24 mmol/L or serum sodium concentration > 145 mmol/L or reduced or temporarily withheld treatment after increase in serum sodium concentration judged by investigator as too rapid

• Treatment‐specific side‐effects

  • AKI: SCr > 1.6 mg/dL

  • Hypotension: not otherwise defined

  • Serious adverse events: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Astellas Pharma US, Inc.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'Eligible patients were stratified by volume status and assigned randomly in a 1:1:1 ratio...'

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'This double‐blind, placebo‐controlled ... study', and 'Fluid intake was limited to ..up to a maximum of 2.0 litres in 24h. ....and 24‐h sodium intake remained stable throughout the study'.

Comment: blinding attempted, and although participants and personnel are unlikely to be fully blinded due to important increases in urine output, all measured outcomes short‐term and fairly objective. Fluid restriction and stable sodium intake were prescribed in all groups

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding outcome assessors probably attempted and all measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 9% attrition, 5% exclusions for never having received the drug, 4% due to treatment discontinuation, reasons well‐documented. Missing data dealt with through analytic techniques for repeated measures using the last available measurement before discontinuation. Given <20% attrition and limited opportunity for introducing bias in such a short‐term study ‐ we judged it 'Low risk'

Selective reporting (reporting bias)

Low risk

Comment: no protocol, short term study (≤1 week), secondary outcomes related to serum sodium concentration and rapid increase in serum sodium concentration reported

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • 'J.K.G. was a principal investigator from 1999‐2004 for trials sponsored by Astellas Pharma US, Inc. N.S. is employed by Astellas Pharma US, Inc.'

  • Comment: senior author is an employee of the company commercialising the study medication.

• Sponsorship bias

  • Quote: 'This study was supported by Astellas Pharma US, Inc (formerly Yamanoushi Pharma America, Inc)

Methods

• Design: parallel RCT (phase 2)

• Hyponatraemia subgroup: yes, but randomisation not stratified according to the subgroup

• Recruitment period: 1998 to 1999

• Treatment duration: 25 days

• Follow‐up: 58 days

Participants

• Countries: Argentina; USA

• Number of centres: 30

• Inclusion criteria: outpatients recruitment from university, community and veterans affairs hospitals; ≥ 18 years; serum sodium concentration: not an inclusion criterion for primary study (we included subgroup with serum sodium concentration ≤ 135 mmol/L); CHF with clinical signs of volume overload; stable oral furosemide dose (40 to 240 mg) ≥ 7 days before enrolment

• Exclusion criteria: cardiac surgery < 90 days prior, MI < 60 days prior; sustained ventricular tachycardia, ventricular fibrillation or automatic implantable cardiac defibrillator discharge < 30 days prior; atrial fibrillation with ventricular rate 115 BPM; diastolic BP > 95 mm Hg; diuretics other than furosemide < 14 days prior; NSAIDS or aspirin 700 mg/d; SCr > 3.0 mg/dL or BUN > 60 mg/dL; serum potassium concentration < 3.4 mmol/L; serum digoxin concentration > 2.2 ng/mL; uncontrolled diabetes mellitus; urinary tract obstruction; morbid obesity; history of intrinsic hepatic disease or liver enzymes ≥ 3 times upper limit of normal; any disorder precluding participation or limiting survival; women if breast‐feeding, of childbearing potential, and not using contraception

• Characteristics randomised participants

  • Number: treatment group 1 (20); treatment group 2 (14); treatment group 3 (15); control group (21)

  • Number, % women: not reported

  • Mean age ± SD (years): not reported

  • Mean serum sodium concentration ± SD (mmol/L): treatment group 1 (133 ± 2); treatment group 2 (133 ± 2); treatment group 3 (134 ± 1); control group (133 ± 1)

  • Cause of hyponatraemia: heart failure (100%)

Interventions

Treatment group 1

• Oral tolvaptan: 1 x 60 mg/d

Treatment group 2

• Oral tolvaptan: 1 x 45 mg/d

Treatment group 3

• Oral tolvaptan: 1 x 30 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: no

• Dietary sodium intake: 'during the study patients were asked to follow a no‐salt‐added diet

• All received standard heart failure treatment including diuretics at the discretion of the investigator

Outcomes

• Response in serum sodium concentration: serum sodium concentration > 135 mmol/L at any time during treatment

• Serum sodium concentration (continuous outcome): defined as change from baseline

• Treatment‐specific side‐effects

  • AKI: insufficiently reported to allow contribution to meta‐analysis

  • Thirst: insufficiently reported to allow contribution to meta‐analysis

  • Polyuria: insufficiently reported to allow contribution to meta‐analysis

• Treatment discontinuation: insufficiently reported to allow contribution to meta‐analysis

Notes

• Outcomes insufficiently reported to allow contribution to meta‐analysis

• Funding source: Otsuka Maryland Research Institute

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: '...central random assignment procedure ... by means of an Interactive Voice‐Response System'

Comment: the randomisation procedure produced groups in the overall study that were not entirely balanced in age and sex, not in terms of medical history or characteristics related to the heart failure. The randomisation was not stratified for serum sodium concentration. The large number of groups relative the total number of participants would have made it likely for the groups to be unbalanced in prognostic characteristics. However, the absence of baseline data specifically for the subgroup with hyponatraemia, make it difficult to judge both magnitude and direction of possible bias.

Allocation concealment (selection bias)

Low risk

Quote: '...central random assignment procedure ... by means of an Interactive Voice‐Response System'.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: '...were assigned to receive...or placebo...', and '...a double‐blind, randomised trial'

Additional info provided by sponsor: 'Both OPC‐41061 and placebo tablets were identical in appearance'.

Comment: blinding attempted, although participants are unlikely to be fully blinded due to important increases in urine output. Also, co‐interventions were not reported in detail for the subgroup with hyponatraemia and included standard heart failure treatment left at discretion of investigator and may have differed between groups according to baseline serum sodium concentration. This could have influenced outcome measurement of serum sodium concentrations

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Additional info provided by sponsor: 'Did not have separate outcome assessors in this trial'.

Comment: blinding outcome assessors attempted, main outcome objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

High risk

Comment: 34% attrition for outcomes related to serum sodium concentration measurements in comparison with numbers provided in the study report

Selective reporting (reporting bias)

High risk

Comment: no protocol available, long‐term study (> 1 week), primary outcomes and secondary outcomes not reported in a way that allowed extraction of raw data

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Comment: 'Dr Gheorghiade is a consultant for and has received a grant support from Otsuka Maryland Research Institute (OMRI), Rockville, Md. Dr. Niazi has received grant support from OMRI, and Drs. Ouyang, Czerwiec, Kambayashi, and Orlandi are OMRI employees.'

  • Comment: multiple authors, including the senior author, are employees of the company commercialising

• Sponsorship bias

  • Comment: this study was industry instigated and supported by Otsuka Maryland Research Institute

Methods

• Design: parallel RCT (phase 2)

• Hyponatraemia subgroup: no

• Recruitment period: 1998 to 1999

• Treatment duration: 12 days

• Follow‐up: 65 days

Participants

• Country: USA

• Number of centres: 9

• Inclusion criteria: university and community hospital in‐patients; ≥ 18 years; serum sodium concentration < 135 mmol/L; euvolaemia or hypervolaemia

• Exclusion criteria: acute coronary ischaemic events < 60 days after randomisation; history of sustained ventricular tachycardia or ventricular fibrillation; SCr > 2.8 mg/dL

• Characteristics randomised participants

  • Number: treatment group (17); control group (11)

  • Number, % women: treatment group: (7, 41%); control group (5, 45%)

  • Mean age ± SD (years): treatment group (66 ± 13); control group (67 ± 9)

  • Mean serum sodium concentration ± SD (mmol/L): treatment group (129 ± 3); control group (129 ± 4)

  • Cause of hyponatraemia: SIADH 10 (36%); heart failure 14 (50%); liver cirrhosis 4 (14%) (not reported for individual groups)

Interventions

treatment group

• Oral tolvaptan: 1 x 10 mg/d up to 1 x 60 mg/d

Control group

• Fluid restriction: ≤ 1.2 L/d with changes left to discretion of investigator + 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: see above

• Dietary sodium intake: a no salt added diet was recommended throughout the study

Outcomes

• Death

• Response in serum sodium concentration (categorical outcome):serum sodium concentration ≥ 135 mmol/L or 10% increase from baseline

• Serum sodium concentration (continuous outcome):change from baseline until day 5

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: not otherwise defined

• Treatment‐specific side‐effects

  • Thirst (continuous outcome): measured on a 0 to 100‐point VAS

  • Hypotension: not otherwise defined

Notes

• Funding source: Otsuka Maryland Research Institute

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: '...subjects were randomised...', and 'This study was a multicenter, prospective, randomised.... trial.'

Additional info provided by the sponsor: 'The study was conducted using local (per site) randomisation.'

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: 'This study was a multicenter, prospective, randomised, open‐label.... trial', and ...subjects were randomised ... to tolvaptan or fluid restriction plus placebo'.

Additional info provided by the sponsor: 'The study drug was supplied as either 5 mg, 15 mg OPC‐4 1061 tablets or matching placebo. All tablets were identical in appearance.' and '(investigators were) not concealed as separate instructions were given to investigators for subjects on fluid‐restriction/placebo and those on active therapy.'

Comment: blinding of participants attempted, and although participants are unlikely to have been fully blinded due to important increases in urine output, serum sodium concentration short‐term and generally objective. Personnel were not blinded

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding of outcome assessors not attempted, but main measured outcomes objective. Participants not blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Quote: 'Two of the 17 subjects assigned to the tolvaptan group and 3 of the 11 subjects assigned to the fluid restriction group were withdrawn during the run‐in phase because of protocol violations or because withdrawal criteria were met.'

Comment: overall 17% attrition ‐12% in experimental group, 27% in placebo group ‐ due to protocol violation or because participants never received the study medication. Reasons not documented in detail. Given <20% attrition, short‐term study, we judged risk of bias to be 'Low'

Selective reporting (reporting bias)

High risk

Comment: no protocol, short‐term study (<1 week), secondary outcomes related to rapid increase in serum sodium concentration not reported. Treatment duration 12 days, yet primary outcome measured at 5 days; Thirst measured at last inpatient visit, which could have been anywhere between 5 and 12 days after study initiation and different for both groups

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Comment: Three authors, including the senior author are affiliated to Otsuka Maryland Research Institute, Rockville, Maryland

• Sponsorship bias

  • Quote: 'This trial was funded by Otsuka Maryland Research Institute'.

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 1 year

• Follow‐up: not reported

Participants

• Country: not reported

• Number of centres: not reported

• Inclusion criteria: serum sodium concentration < 130 mmol/L; liver cirrhosis and ascites

• Exclusion criteria: not reported

• Characteristics randomised participants

  • Number: treatment group (92); control group (47)

  • Number, % women: not reported

  • Mean age ± SD (years): not reported

  • Serum sodium concentration (mmol/L): not reported

  • Control group: liver cirrhosis 47 (100%)

Interventions

Treatment group

• Oral satavaptan: 1 x 5 mg/d up to 1 x 50 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: not reported

• Dietary sodium intake: not reported

Outcomes

• Length of hospital stay: average number of days/year/patient

• Treatment‐specific side‐effects

  • AKI: SCr increase > 50% from baseline or SCr > 1.5 mg/dL

Notes

• Funding source: not reported

• Publication: abstract only

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'Patients...were randomised...'.

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'. No baseline characteristics presented

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: 'Patients...were randomised into a multi‐centre, single‐blind study to receive in a 2:1 ratio either a flexible dose‐regimen of satavaptan (5‐50 mg once daily) or placebo'.

Comment: blinding of participants probably attempted. However, insufficient information for assessing to what extent it was achieved, or to permit judgement of 'High risk' or 'Low risk'. Blinding of personnel not attempted. No information on how differential treatment and follow‐up was avoided

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Quote: 'Patients...were randomised into a multi‐centre, single‐blind study...'

Comment: outcome assessors not blinded and main outcome measure ‐ objective outcome measures

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Selective reporting (reporting bias)

High risk

Comment: no protocol and long term (> 1 week) study, primary outcomes and secondary outcomes related to rapid increase in serum sodium concentration not reported

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Comment: no declaration of interest. From declarations of interest from Gines 2008a it is clear that 'Dr. Ginès is a consultant for Sanofi‐Aventis. Dr. Wong is a consultant for Sanofi‐Aventis. Dr. Watson owns stock in Sanofi‐Aventis'

• Sponsorship bias

  • Comment: all of the studies in which a vasopressin receptor antagonist was evaluated were industry sponsored. It seems very likely this one was industry sponsored too, but at this stage insufficient information

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: 2009 to 2010

• Treatment duration: 6 months

• Follow‐up: 7 months

Participants

• Countries: Belgium; Czech Republic; Israel; India; Italy; Mexico; Peru; USA

• Number of centres: 61

• Inclusion criteria: long‐term care facilities and nursing homes; ≥ 18 years; serum sodium concentration < 135 mmol/L; euvolaemia; willing to be observed in a monitored setting for the first 8 hours following treatment initiation; in the Investigator's judgement the patient has adequate visual and auditory acuity to allow participation in the trial

• Exclusion criteria: overt symptoms requiring immediate intervention; acute or transient hyponatraemia, pseudohyponatraemia, hypovolaemic hyponatraemia; drug‐induced hyponatraemia; hyponatraemia due to hypothyroidism or adrenal insufficiency; hypertonic hyponatraemia, hypokalaemia; demeclocycline, lithium carbonate, urea, or any other vasopressin antagonist within 7 days of participation; supine systolic BP ≤ 90 mm Hg; SCr > 3 mg/dL; uncontrolled diabetes; severe malnutrition; documented cirrhosis or alcoholic liver disease; terminal illness; psychogenic polydipsia; urinary tract obstruction; myocardial ischaemia; MI or cerebrovascular accident within 30 days; neurologic impairment such as dementia; conditions causing inability to respond to thirst; NYHA class III or IV; women pregnant or breastfeeding or planning to

• Characteristics randomised participants

  • Number: treatment group (154); control group (52)

  • Number, % women: treatment group (81, 53%); control group (25, 48%)

  • Mean age ± SD (years): treatment group (67 ± 14); control group (63 ± 14)

  • Mean serum sodium concentration ± SD (mmol/L): treatment group (130 ± 4); control group (130 ± 4)

  • Cause of hyponatraemia: SIADH (100%)

Interventions

Treatment group

• Oral lixivaptan: 1 x 25 mg/d, titrated to 1 x 100 mg/d

Control group

• Oral matching placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: investigators were instructed, when possible, to not institute fluid restriction for at least the first 72 hours of dosing. However, fluid restriction could be initiated at the investigator's discretion throughout the duration of the study

• Dietary sodium intake: not reported

Outcomes

• Death

• Response: serum sodium concentration 135 to ≤ 145 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Cognitive function (continuous outcome): measured by the time to complete the Trail Making Test ‐ Part B; continuous outcome ‐ as measured by change from baseline in the Medical Outcomes Study 6‐Item Cognitive Function Scale (MOS‐6) score ‐ insufficiently reported to allow contribution to meta‐analysis

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: serum sodium concentration ≥ 145 mmol/L

  • Rapid increase in serum sodium concentration: serum sodium concentration increase ≥ 8 mmol/L during first 8 hours, > 12 mmol/L during first 24 hours or > 18 mmol/L during first 48 hours

  • Osmotic demyelination syndrome: not otherwise defined

• Treatment‐specific side‐effects

  • Hypotension: not otherwise defined

  • Thirst: not otherwise defined

  • Polyuria: not otherwise defined

  • Adverse event: treatment‐emergent adverse events, not otherwise defined

  • Serious adverse events: serious treatment‐emergent adverse events, not otherwise defined

• Treatment discontinuation

Notes

• Funding source: No clear statement about funding but from author affiliations and declarations of interest clear this is an industry instigated study

• Additional data source: FDA Briefing Document for the Cardiovascular and Renal Drugs Advisory (CRDAC)

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'The HARMONY study was a ... randomised .... study', and 'Subjects were initially randomised to ... '.

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'The HARMONY study was a ... double‐blind, placebo‐controlled ... study'.

Comment: blinding probably attempted, although due to increase in urine output, un‐blinding likely to have occurred. Fluid restriction was largely left to the discretion of the investigator, but fewer participants receiving the study medication had fluid restriction initiated or tightened by day 30 (11% versus 21%)

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

High risk

Comment: overall 17% discontinuation at 8 weeks, numbers and reasons similarly distributed between experimental and placebo group. However, 67% discontinuation at 24 weeks, according to the FDA briefing document because, quote 'According to the applicant, because the primary and secondary efficacy endpoints targeted time points between the initial eight weeks from randomisation, therefore, when the last subject enrolled in the study had completed eight weeks of treatment, the applicant instructed the investigators to stop blinded therapy in all subjects in the study'

Selective reporting (reporting bias)

Low risk

Comment: the study was registered with ClinicalTrials.gov, but no clear listing of predefined outcome time‐points. That said, all preregistered outcomes and outcomes related to over correction reported. Some additional outcomes were insufficiently reported for contribution to meta‐analysis

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'WTA has received consulting fees from Cardiokine. GD has received consulting fees from Wyeth (Pfizer), Sonofi, Cardiokine, Otsuka Pharmaceuticals, and Yamanouchi. RCJ has served as a Principal Investigator on vaptan‐related phase III clinical trials sponsored by Astellas, Otsuka, and Cardiokine. He has also served as a consultant for Otsuka and Cardiokine. NK has served on the speakers' bureau for Otsuka. DGB has received consulting fees from Cardiokine, received consulting fees from Cardiokine, receives grants from Otsuka, and is a paid consultant for Otsuka Canada. CO is an employee of Cardiokine. MM received consulting fees from Otsuka. HPT and YY have no relevant relationships to disclose.'

• Sponsorship bias

  • Quote: 'Medical writing and editorial assistance, supported by Cardiokine, ...'

  • No clear statement about funding but from author affiliations and declarations of interest clear this is an industry instigated study

Methods

• Design: parallel RCT

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: not reported

• Follow‐up: not reported

Participants

• Country: UK

• Number of centres: not reported, presumably 1

• Inclusion criteria: serum sodium concentration < 130 mmol/L; liver cirrhosis and ascites

• Exclusion criteria: not reported

• Characteristics randomised participants

  • Number: 10 (not reported for individual groups)

  • Number, % women: not reported

  • Mean age ± SD (years): not reported

  • Serum sodium concentration (mmol/L): not reported

  • Cause of hyponatraemia: liver cirrhosis (100%)

Interventions

Treatment group

• Oral urea: 3 x 20 mg/d within 24 hours of development of hyponatraemia

Control group

• Oral urea: 3 x 20 mg/d after 3 days of no treatment after development of hyponatraemia

Co‐interventions (all groups)

• Fluid restriction: not reported

• Dietary sodium intake: not reported

Outcomes

• Serum sodium concentration (mmol/L) from beginning to end of treatment (no comparative data)

• Treatment‐specific side‐effects

  • Diarrhoea and vomiting: no data provided

Notes

• Funding source: not reported

• Publication: abstract only

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'Patients...randomised...'.

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'. No baseline characteristics presented.

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Selective reporting (reporting bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Other bias

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'.

Methods

• Design: parallel RCT (phase 2)

• Hyponatraemia subgroup: no

• Recruitment period: 2004 to 2005

• Treatment duration: 14 days

• Follow‐up: 21 days

Participants

• Countries: Argentina; Australia; Belgium; Canada; Croatia; Czech Republic; France; Germany; Hungary; Italy; Romania; Spain; USA

• Number of centres: 30

• Inclusion criteria: ≥ 18 years; serum sodium concentration ≤ 130 mmol/L; cirrhosis diagnosed by liver biopsy or combination of clinical, biochemical, ultrasonographic and endoscopic findings; moderate/tense ascites defined as grade 2‐3 ascites according to previously established criteria

• Exclusion criteria: serum bilirubin concentration > 8 mg/dL; INR > 3; platelet count < 30,000/mm³; SCr > 2mg/dL; serum potassium concentration > 5.5 mmol/L; hepatocellular carcinoma > Milan criteria; hepatic encephalopathy > grade 1; bacterial infection/gastrointestinal bleeding < 10 days; large‐volume paracentesis < 7 days prior; cardiac disease, systolic BP < 80 mm Hg; inducer or inhibitors of CYP450 CYP3A < 14 days prior

• Characteristics randomised participants

  • Number: treatment group 1 (28); treatment group 2 (26); treatment group 3 (28); control group (28)

  • Number % women: treatment group 1 (8, 29%); treatment group 2 (7, 27%); treatment group 3 (12, 43%); control group (6, 21%)

  • Mean age ± SD (years): treatment group 1 (59 ± 10); treatment group 2 (56 ± 9); treatment group 3 (57 ± 8); control group (55 ± 10)

  • Mean serum sodium concentration ± SD (mmol/L): treatment group 1 (126 ± 6); treatment group 2 (128 ± 4); treatment group 3 (127 ± 5); control group (126 ± 4)

  • Cause of hyponatraemia: liver cirrhosis (100%)

Interventions

Treatment group 1

• Oral satavaptan: 1 x 25 mg/d

Treatment group 2

• Oral satavaptan: 1 x 12.5 mg/d

Treatment group 3

• Oral satavaptan: 1 x 5 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: < 1.5 L/d, unless patients were thirsty or there was a medical need to increase fluid intake

• Dietary sodium intake: ≤ 2.7 g/d

Outcomes

• Response in serum sodium concentration (categorical outcome): increase from baseline in serum sodium concentration ≥ 5 mmol/L or serum sodium concentration ≥ 135 mmol/L at any time during treatment

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Outcomes related to over‐correction of serum sodium concentration

  • Rapid increase in serum sodium concentration: serum sodium concentration increase ≥ 8 mmol/L/24 h

  • Hypernatraemia: serum sodium concentration ≥ 145 mmol/L at any point during treatment

  • Osmotic demyelination syndrome: neurological symptoms

• Treatment‐specific side‐effects

  • Thirst (categorical outcome): measured on a VAS (cut‐offs for categorisation not reported)

  • Adverse events: not otherwise defined

  • Serious adverse events: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Sanofi‐Aventis

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: 'Each participating centre received blocks of study treatment which contained equal numbers of each of the four therapies on a random sequence generated by a computer program', and 'The median number of patients included in the 30 participating centres was 3 (range 1 to 14)'.

Comment: The randomisation procedure generated groups well‐balanced for age, sex and baseline serum sodium concentration. The severity of liver disease as assessed by the serum bilirubin concentration, Child‐Pugh and MELD score was greater for the placebo group in comparison with the other groups. We assumed the influence of the imbalance thus generated would be negligible in practice

Allocation concealment (selection bias)

Low risk

Comment: considering randomisation procedure and extent of attempting to blind, likely adequate

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'This was a ...double‐blind, placebo‐controlled efficacy study', and 'all...were maintained on a low‐sodium diet ... and daily fluid intake was limited to 1.5L unless patients were thirsty or there was a medical need to increase fluid intake'

Comment: blinding attempted, and although participants and personnel are unlikely to be fully blinded due to important increases in urine output, all measured outcomes short‐term and fairly objective. Fluid restriction occurred in all groups, any increase in fluid intake likely to have occurred in the intervention groups, with possible bias favouring the placebo group for outcomes related to serum sodium concentration

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding outcome assessors probably attempted, main measured outcomes objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 5% attrition, presumably due to treatment discontinuation, despite efforts to measure outcomes regardless. Reasons for treatment discontinuation well documented. Missing data dealt with through analytic techniques for repeated measures using the last available measurement before discontinuation. Given <20% attrition and limited opportunity for introducing bias in such a short‐term study ‐ we judged it 'Low risk'

Selective reporting (reporting bias)

High risk

Comment: protocol registered after completion of the study. Outcomes related to serum sodium concentration said to be measured at 14 days, yet only reported at day 5. Protocol also lists 'trail making test' and 'quality of life' as outcomes, which are not reported in the study publication

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'Dr. Ginès is a consultant for Sanofi‐Aventis. Dr. Wong is a consultant for Sanofi‐Aventis. Dr. Watson owns stock in Sanofi‐Aventis.'

• Sponsorship bias

  • Quote: 'The sponsor, Sanofi‐Aventis, and the academic principal investigator designed the study, developed the protocol, and prepared the first and subsequent drafts of the manuscript', and 'The sponsor and the academic principal investigator held and analysed the data'

  • Comment: the principal investigator is a consultant for the company developing the study drug

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: 2007 to 2009

• Treatment duration: 21 days

• Follow‐up: 28 days

Participants

• Country: USA

• Number of centres: 12

• Inclusion criteria: ≥ 50 years; serum sodium concentration 123 to ≤ 134 mmol/L and for 118 to ≤ 122 mmol/L inclusion based on consultation and approval from the study medical monitor

• Exclusion criteria: conditions or history which may present a safety concern to the subject or their offspring or extreme susceptibility to hypotension with sudden fluid loss; hyponatraemia that is acute, easily reversible, artifactual, or due to a condition not associated with vasopressin excess; conditions associated with an independent imminent risk of morbidity and mortality; conditions which may confound the assessment of endpoints, history of poor compliance, participation in a clinical trial believed by the principal investigator or sponsor likely to confound endpoint assessments; conditions which may confound primary endpoints of cognitive function

• Characteristics randomised participants

  • Number: treatment group (29); control group (28)

  • Number, % women: treatment group (15, 52%); control group (19, 68%)

  • Mean age ± SD (years): treatment group (71 ± 10); control group (71 ± 10)

  • Serum sodium concentration (mmol/L): not reported

  • Cause of hyponatraemia: not reported

Interventions

Treatment group

• Oral tolvaptan: 1 x 15 mg/d up to 1 x 60 mg/d

Control group

• Oral matching placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: fluid restrictions to be loosened during dose titration phase, reinstated at any time if serum sodium concentration fails to improve, left at discretion of investigator

• Dietary sodium intake: not reported

Outcomes

• Serum sodium concentration (continuous outcome): change from baseline until end of treatment

• Cognitive function (Neurocognitive Composite Score of Speed Domains) change from baseline in sum of all speed domain Z‐scores: reaction time (Simple = recognize "yes" 50 times; choice = recognize "yes" or "no" 50 times; digit vigilance = match 45 digits); psychomotor speed (morse tapping = tap button for 30 s with right & left hands); processing speed (rapid visual information processing = detect consecutive sequences of 3 odd or 3 even digits; numeric working memory = recognize numbers from series of 5 digits among 30; word recognition = remember 15 prior learned words from 30 total; results age‐matched to healthy controls from Cognitive Drug Research normative data

• Gait test: change from baseline timed get‐up‐and‐go test = time it takes for a seated subject to rise from a chair, walk 3 m, walk around an object and return to sit in chair. Values: under 10 s (no difficulties), 10 to 20 s (starting to have balance difficulty), over 30 s (at high risk for falls and dependent in most activities of daily living and mobility); test assesses risk to elderly subjects of falling and higher scores in seconds indicate higher risk of falling

• Treatment‐specific side‐effects

  • Thirst: not otherwise defined

  • Adverse event: not otherwise defined

  • Serious adverse events: serious treatment‐emergent adverse events

• Treatment discontinuation

Notes

• Funding source: Otsuka Pharmaceutical Development & Commercialization

• Publication: protocol registry only

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'Subjects will be randomised...'

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: 'A pilot...double‐blind, placebo‐controlled ... study', and 'Subjects will be randomised to ... or matching placebo tablet', and 'during this period, fluid restrictions should be loosened or suspended, until the subject's response to therapy can be evaluated, typically over the first few days of therapy. Fluid restriction may be reinstated at any time in subjects whose sodium fails to improve or worsens with study therapy. Subjects entering ...with a serum sodium concentration less than 130 mEq/L may be fluid restricted if necessary at the discretion of the investigator'.

Comment: blinding of both participants and personnel attempted. However, unlikely for participants to have been fully blinded due to likely important increases in urine output. Also, fluid restriction was not mandatory and could be adapted at the discretion of the investigator. This may have biased outcome measures at one month in favour of the study medication

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective. Unlikely for participants to have been fully blinded, and plausible 'high risk' of bias for self‐reported outcome of thirst

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 9% attrition, 8% due to treatment discontinuation, asymmetrical due to adverse events in the experimental group. Given < 10% attrition in long‐term study, we judged opportunity for introducing attrition bias 'Low'

Selective reporting (reporting bias)

Low risk

Comment: study registered with ClinicalTrials.gov in 2007 before study initiation. Primary and secondary outcomes including time‐points match

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • No declaration of interest

• Sponsorship bias

  • Quote: 'Sponsor: Otsuka Pharmaceutical Development & Commercialization'

  • Comment: no further information as study not published at this time

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 7 days

• Follow‐up: not reported

Participants

• Country: UK

• Number of centres: not reported (presumably 1)

• Inclusion criteria: serum sodium concentration < 130 mmol/L despite 3 days of fluid restriction; liver cirrhosis with refractory ascites

• Exclusion criteria: active infection; hepatic encephalopathy > grade 2; creatinine concentration >150 µmol/L; systolic BP < 80 mm Hg; alcoholic hepatitis; adrenal insufficiency; hypothyroidism; hyperthyroidism; diabetes; severe cardiovascular disease; head injury; diuretics within 7 days; paracentesis < 1 week

• Characteristics randomised participants

  • Number: 24

  • Number, % women: not reported

  • Mean age ± SD (years): not reported

  • Serum sodium concentration (mmol/L): not reported

  • Cause of hyponatraemia: not reported; liver cirrhosis (100%)

Interventions

Treatment group

• IV human salt‐poor albumin: 1 x 40 g/d

Control group

• No treatment

Co‐interventions (all groups)

• Fluid restriction: < 1.5 L/d

• Dietary sodium intake: 1.8 g/d

Outcomes

• Serum sodium concentration (continuous outcome): absolute values; insufficiently reported to allow contribution to meta‐analysis

Notes

• Funding source: not reported

• Publication: abstract only

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: '...a randomised clinical trial'.

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Quote: 'treatment group 1: Fluid restriction (1.5 L)+Na restriction (80 mmol/d), Group 2: Human Salt Poor Albumin (40 gm/d) + Fluid (1.5 L) + Na restriction (80 mmol/d)

Comment: attempts at blinding not mentioned, blinding unlikely to have happened

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Selective reporting (reporting bias)

High risk

Comment: no protocol, short‐term study (≤ 1 week) but no reporting of primary outcomes death or health‐related quality of life, and no reporting of secondary outcomes related to rapid increase in serum sodium concentration in a way that permitted data extraction

Other bias

Unclear risk

• Bias through possible financial conflict of interest of the authors

  • No declaration of interest

• Sponsorship bias

  • Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 2 days

• Follow‐up: 7 days

Participants

• Countries: India; Israel; USA

• Number of centres: 25

• Inclusion criteria: ≥ 18 years; serum sodium concentration 115 to ≤ 134 mmol/L; euvolaemia or hypervolaemia

• Exclusion criteria: women of childbearing potential with positive pregnancy test; clinical evidence of volume depletion; mechanical circulatory or respiratory support; supine systolic BP < 85 mm Hg; history of significant orthostatic hypotension; requiring emergent treatment for hyponatraemia

• Characteristics included participants

  • Number: treatment group 1 (29); treatment group 2 (30); treatment group 3 (30); control group (28)

  • Number, % women: treatment group 1 (18, 62%); treatment group 2 (13, 43%); treatment group 3 (16, 53%); control group (14, 40%)

  • Mean age ± SD (years): treatment group 1 (63 ± 22); treatment group 2 (61 ± 20); treatment group 3 (65 ± 14); control group (59 ± 20)

  • Serum sodium concentration (mmol/L): treatment group 1 (124 ± 5); treatment group 2 (124 ± 4); treatment group 3 (125 ± 4); control group (125 ± 4)

  • Cause of hyponatraemia

    • Treatment group 1: not specified; euvolaemia (27, 93%); hypervolaemia (2, 7%)

    • Treatment group 2: not specified; euvolaemia (28, 93%); hypervolaemia (2, 7%)

    • Treatment group 3: not specified; euvolaemia (27, 90%); hypervolaemia (3, 10%)

    • Control group: not specified; euvolaemia (25, 89%); hypervolaemia (3, 11%)

Interventions

Treatment group 1

• IV conivaptan: 20 mg loading dose in 30‐min infusion + conivaptan 1 x 20 mg/d in premixed bag without propylene glycol solvent, in continuous infusion

Treatment group 2

• IV conivaptan: 20 mg loading dose in 30‐min infusion + conivaptan 1 x 20 mg/d in ampoule with propylene glycol solvent, in continuous infusion

Treatment group 3

• IV placebo: loading dose in 30‐min infusion + conivaptan 1 x 20 mg/d in premixed bag without propylene glycol solvent, in continuous infusion

Control group

• IV placebo: loading dose in 30‐min infusion + conivaptan 1 x 20 mg/d in ampoule with propylene glycol solvent, in continuous infusion

Co‐interventions (all groups)

• Fluid restriction: < 2 L/d

• Dietary sodium intake: not reported

Outcomes

• Death: time‐point measurement unclear

• Response in serum sodium concentration (categorical outcome): increase from baseline in serum sodium concentration > 4 mmol/L at any time during treatment

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Outcomes related to over‐correction of serum sodium concentration

  • Rapid increase in serum sodium concentration: serum sodium concentration increase > 12 mmol/L during the first day or serum sodium concentration > 145 mmol/L at any time during treatment

• Treatment‐specific side‐effects

  • Adverse events: not otherwise defined

  • Serious adverse events: not otherwise defined

  • Injection‐site reactions (continuous outcome): severity of injection‐site reactions as measured by the ISR modified 5‐point scale; and as categorical outcome not otherwise defined

  • Injection‐site phlebitis: not otherwise defined

  • Injection‐site thrombosis: not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Astellas Pharma US, Inc.

• Request for additional information sent 27/10/2014

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'This was a randomised, parallel‐group study..', and 'Eligible patients were... randomised to...'

Comment: large number of groups relative to the number of included participants may have caused simple randomisation to fail: groups are not entirely balanced in terms of age and sex. Insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'The placebo ampoules and premixed bags contained all the components of the active formulation except conivaptan. For the ampoule formulation, the contents of the 4 mL ampoule was added to a bag containing 96 mL of 5% dextrose injection, so that the final volume was 100 mL for the loading dose or a bag containing 246 mL of 5% dextrose injection for the continuous infusion.'

Comment: blinding of participants and personnel attempted and given all groups included active ingredient, un‐blinding due to differences in urine output not very likely

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding outcome assessors adequate

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 17% attrition due to exclusion after randomisation (3%) and treatment discontinuation (14%); reasons well‐documented. Missing data dealt with through analytic techniques for repeated measures using the last available measurement before discontinuation. Given < 20% attrition and limited opportunity for introducing bias in such a short‐term study ‐ we judged it 'Low risk'

Selective reporting (reporting bias)

Low risk

Comment: no protocol and short term study (< 1 week), secondary outcomes related to both serum sodium concentration and rapid increase in serum sodium correction reported. Because a repeated measures analytic technique was used, unclear at what time point serum sodium concentration was measured for all participants, yet given short term study, bias thus introduced probably negligible in practice

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'Dr. Velez has received a grant and honorarium from and served on the speakers bureau for Astellas Pharma. Drs McNutt and Abeyratne and Ms. Klasen are employed by Astellas Pharma US, Inc.'

  • Comment: Senior authors are employed by company commercialising the study medications

• Sponsorship bias

  • Comment: this study was sponsored by Astellas Pharma, Inc

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 2 days

• Follow‐up: 30 days

Participants

• Countries: India; USA

• Number of centres: 16

• Inclusion criteria: hospital in‐patients; ≥ 18 years; serum sodium concentration 115 to ≤1 30 mmol/L; euvolaemia or hypervolaemia

• Exclusion criteria: women of childbearing potential with positive pregnancy test; volume depletion or dehydration; haemodynamic instability; unable to provide consent due to significantly altered mental status; hypertonic saline administered in previous 24 h; expected to require emergent treatment of hyponatraemia supine systolic BP < 85 mm Hg; significant orthostatic hypotension; uncontrolled hypertension; uncontrolled arrhythmias requiring emergent pacemaker placement or treatment; severe hypothyroidism; hyperthyroidism; adrenal insufficiency; creatine clearance < 20 mL/min, urinary outflow obstruction unless catheterized; alanine transaminase or aspartate transaminase > 5 times upper limit of normal, serum albumin concentration ≤ 15 g/L; white blood cell count < 3 x 10⁹/L, concomitant illness that could interfere with treatment or evaluation of safety; history of hepatic encephalopathy, hematemesis or melena, moderate‐to‐severe hepatic impairment, moderate ascites secondary to hepatic dysfunction, non‐fasting blood glucose concentration ≥ 200 mg/dL

• Characteristics included participants

  • Number: treatment group 1 (20); treatment group 2 (20); control group (9)

  • Number, % women: treatment group 1 (11, 55%); treatment group 2 (7, 35%); control group (8, 89%)

  • Mean age ± SD (years): treatment group 1 (68 ± 15); treatment group 2 (59 ± 17); control group (62 ± 20)

  • Serum sodium concentration (mmol/L): treatment group 1 (126 ± 4); treatment group 2 (126 ± 4); control group (126 ± 4)

  • Cause of hyponatraemia

    • Treatment group 1: SIADH (9, 45%); heart failure (6, 30%); unclear (5, 25%)

    • Treatment group 2: SIADH (6, 30%); heart failure (5, 25%); nephrotic syndrome (1, 5%); alcoholism (1, 5%); unclear (6, 30%)

    • Control group: SIADH (2, 22%); heart failure (3, 33%); impaired salt intake due to hypertension (1, 11%); unclear (3, 33%)

Interventions

Treatment group 1

• IV conivaptan: 2 x 20 mg/d in 30‐min infusion

Treatment group 2

• IV conivaptan: 1 x 20 mg/d in 30‐min infusion

Control group

• IV placebo: 2 x 100 mL 5% dextrose in water in 30‐min infusion

Co‐interventions (all groups)

• Fluid restriction: < 2 L/d

• Dietary sodium intake: not reported

Outcomes

• Death

• Response in serum sodium concentration (categorical outcome): increase from baseline in serum sodium concentration > 4 mmol/L at any time during treatment

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: serum sodium concentration > 145 mmol/L; insufficiently reported to allow contribution to meta‐analysis

  • Rapid increase in serum sodium concentration: serum sodium concentration increase > 12 mmol/L during the first day

  • Osmotic demyelination syndrome: not otherwise specified

• Treatment‐specific side‐effects

  • Hypotension: supine systolic BP < 85 mm Hg

  • Adverse events: not otherwise defined

  • Serious adverse events: not otherwise defined

  • Injection‐site reactions

  • Injection‐site phlebitis

  • Injection‐site thrombosis

• Treatment discontinuation

Notes

• Funding source: Astellas Pharma

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'This randomised, double‐blind...' and 'Eligible patients were stratified by baseline SSC (serum sodium concentration) ... and then randomly assigned...

Comment: methods for sequence generation not reported. Due to the large number of groups relative to the number of included participants, adequate simple randomisation may have failed to produce groups with similar baseline prognosis. E.g. there is an imbalance in age, sex and volume status across the groups. We assumed the influence of the imbalance thus generated would be negligible in practice

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'In each dosing regimen, conivaptan or placebo was administered via a 30‐minute i.v. infusion using a 100 mL plastic bag containing either 20 mg conivaptan hydrochloride in 5% dextrose injection or placebo (5% dextrose injection). In the once‐daily regimen, conivaptan was administered at 0 and 24 hours, and placebo was administered at 12 and 36 hours to maintain blinding'

Comment: blinding attempted, and although participants and personnel are unlikely to be fully blinded due to important increases in urine output, all measured outcomes very short‐term and fairly objective

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding outcome assessors probably attempted and all measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: overall 18% attrition due to exclusion after randomisation (2%) and treatment discontinuation (16%); reasons fairly well‐documented. All save 1 were re‐included in the analysis for sodium concentration measurements using the analytic techniques for repeated measures using the last available measurement before discontinuation. Given < 20% attrition and limited opportunity for introducing bias in such a short term study, we judged it 'Low risk of bias'

Selective reporting (reporting bias)

Low risk

Comment: no protocol and short term study (≤ 1 week), secondary outcomes related to both serum sodium concentration and rapid increase in serum sodium correction reported

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'Ms. Klasen, Dr Abeyratne, and Dr. McNutt are employees of Astellas Pharma US Inc.'

  • Comment: Several authors are employees of the company who commercialised the study drug

• Sponsorship bias

  • Quote: 'Supported by a grant from Astellas Pharma Global Development, Inc.'

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: 2008 to 2010

• Treatment duration: 30 days

• Follow‐up: 60 days

Participants

• Country: Belgium; Canada; Germany; India; Poland; USA

• Number of centres: 37

• Inclusion criteria: ≥ 18 years; serum sodium concentration 120 to < 130 mmol/L; euvolaemia; hospitalised or willing to stay for 48 to 72 hours

• Exclusion criteria: overt symptoms requiring immediate intervention; acute or transient hyponatraemia, pseudohyponatraemia or hypertonic hyponatraemia; hyponatraemia due to hypothyroidism, adrenal insufficiency; psychogenic polydipsia; creatinine > 3 mg/dL, HbA1c > 9%, urinary tract obstruction; severe pulmonary artery hypertension expected to deteriorate, NYHA III or IV; ST‐segment elevation, MI previous 30 days, active myocardial ischaemia; cerebrovascular accident previous 30 days, significant neurological impairment; nephrotic syndrome, advanced liver disease, documented cirrhosis or alcoholic liver disease; terminal illness; radiotherapy or chemotherapy within 2 weeks; demeclocycline, lithium carbonate, urea, vasopressin antagonist within 7 days; women pregnant or breastfeeding or planning to

• Characteristics randomised participants

  • Number: treatment group (54); control group (52)

  • Number, % women: treatment group 1 (28, 52%); control group (22, 42%)

  • Mean age ± SD (years): treatment group (66 ± 14); control group (65 ± 13)

  • Serum sodium concentration (mmol/L): treatment group (125 ± 5); control group (124 ± 5)

  • Cause of hyponatraemia: SIADH (100%)

Interventions

Treatment group 1

• Oral lixivaptan: 1 x 50 mg/d, titrated to 1 x 25 mg/d or 1 x 100 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: investigators were instructed, when possible, to not institute fluid restriction for at least the first 72 hours of dosing. However, fluid restriction could be initiated at the investigator's discretion throughout the duration of the study

• Dietary sodium intake: not reported

Outcomes

• Death

• Response: serum sodium concentration ≥ 135 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline until treatment discontinuation

• Cognitive function (continuous outcome): time to complete the Trail Making Test ‐ Part B; (continuous outcome) change from baseline in the Medical Outcomes Study 6‐Item Cognitive Function Scale (MOS‐6) score; insufficiently reported to allow contribution to meta‐analysis

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: serum sodium concentration ≥ 145 mmol/L

  • Rapid increase in serum sodium concentration: serum sodium concentration increase ≥ 8 mmol/L during first 8 h, > 12 mmol/L during first 24 hours or > 18 mmol/L during first 48 h

  • ODS: clinical signs and symptoms consistent with the diagnosis

• Treatment‐specific side‐effects

  • Hypotension: not otherwise defined

  • Adverse event: treatment‐emergent adverse events, not otherwise defined

  • Serious adverse events: serious treatment‐emergent adverse events, not otherwise defined

• Treatment discontinuation

Notes

• Funding source: Cardiokine Biopharma.

• Additional data source: FDA Briefing Document for the Cardiovascular and Renal Drugs Advisory (CRDAC)

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'The LIBRA study was a ... randomised .... study', and 'Subjects were initially randomised to ... '

Comment: Overall insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'The Libra study was a ... double‐blind, placebo‐controlled ... study'

Comment: blinding probably attempted, although due to increase in urine output, un‐blinding likely to have occurred. Fluid restriction was largely left to the discretion of the investigator, and more participants receiving the study medication had fluid restriction initiated or tightened by day 30 (32% versus 23%), but more participants in the placebo group were on fluid restriction at baseline (37% versus 65%) and at any time‐point throughout the study

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: blinding insufficiently reported but main outcomes objective such that un‐blinding of outcome detection unlikely to be an important source of bias

Incomplete outcome data (attrition bias)
All outcomes

High risk

Comment: overall 13% discontinuation at 7 days, 5% of which never received treatment; 67% of discontinuations occurred in placebo group. 31% discontinuation at 30 days, with last observation carried forward method as imputation method for dealing with missing serum sodium concentration values, which may have caused overestimation of the effect of the study medication. Cognitive function tests were taken at day 30 or the early termination visit. We judged outcomes measured at 7 days to at 'Low risk', but those at 30 days to be at 'High risk' of attrition bias

Selective reporting (reporting bias)

Low risk

Comment: protocol first registered with ClinicalTrials.gov in 2008. Primary registered outcome was average daily area under the curve (AUC) of change from baseline in serum sodium concentrations up to 72 hrs, whereas the reported primary outcome is change in serum sodium concentration at 7 days. That said, all reregistered outcomes were reported and secondary outcomes related to over‐correction also reported

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'WTA has received consulting fees from Cardiokine. PA.... has received honoraria for this work and from Otsuka for presentations and participation in advisory committees. DGB has received consulting fees from Cardiokine and received grants from and is paid consultant for Otsuka Pharmaceuticals. RCJ has served as a principal investigator on vaptan‐related phase‐III clinical trials sponsored by Astellas, Otsuka, and Cardiokine. He has also served as a consultant for Otsuka and Cardiokine. At the time of the study and manuscript preparation, CO was an employee of Cardiokine Biopharma'

• Sponsorship bias

  • Quote: 'Medical writing and editorial assistance, supported by Cardiokine Biopharma, ...'

  • No clear statement about funding but from author affiliations and declarations of interest clear this is an industry instigated study

Methods

• Design: parallel RCT (phase 4)

• Hyponatraemia subgroup: no

• Recruitment period: 2008 to 2009

• Treatment duration: 1 day

• Follow‐up: until discharge from hospital

Participants

• Country: USA

• Number of centres: 1

• Inclusion criteria: patients admitted to a dedicated neuro/spine intensive care unit of a teaching hospital; ≥ 18 years: serum sodium concentration < 135 mmol/L and Glasgow Coma Scale < 15; or < 130 mmol/L

• Exclusion criteria: enrolment in the NMH high‐risk spine protocol; expected death from any cause; known sensitivity or allergy to conivaptan; baseline creatinine > 1.5 mg/dL; clinical diagnosis of hypovolaemia, or by central venous pressure < 5 mm Hg if a central venous catheter is in place; concomitant use of potent inhibitors of cytochrome P‐450 isoenzyme 3A4; clinical diagnosis of liver failure or insufficiency; pregnancy; intra‐arterial vasodilators within 24 h; concern by the Neuro‐intensive care unit pharmacist of a drug‐drug interaction that would meaningfully impact care; diabetes insipidus or treated with vasopressin; congestive heart failure, inclusion declined by the attending physician or consulting study nephrologist

• Characteristics randomised participants

  • Number: treatment group (3); control group (3)

  • Number, % women: not reported

  • Median age, range (years): treatment group (20, 45 to 64); control group (65, 67 to 69)

  • Mean serum sodium concentration (mmol/L): treatment group (131); control group (132)

  • Cause of hyponatraemia

    • Treatment group 1: SIADH (100%) (central nervous system‐infection (1); brain metastasis (1); tuberous sclerosis (1))

    • Control group: SIADH (100%) (central nervous system‐infection (1); seizures (1); cerebrotrauma (1))

Interventions

Treatment group

• IV conivaptan: 1 x 20 mg loading dose + 20 mg/d continuous infusion

Control group

• Usual care

Co‐interventions (all groups)

• Fluid restriction: at discretion of treating physician

• sodium intake: at discretion of treating physician

Outcomes

• Death

• Serum sodium concentration (continuous outcome): change from baseline

• Treatment‐specific side‐effects

  • Hypotension: systolic BP < 100 mm Hg or new vasopressor medication

Notes

• Funding: Astellas Pharma

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: '...the patient was randomised by opening the lowest numbered, sealed envelope.'

Comment: although methods for generating the sequence were not reported, we judged it likely for methods to have been adequate given researchers went to trouble of using sealed, numbered envelopes

Allocation concealment (selection bias)

Low risk

Quote: '...the patient was randomised by opening the lowest numbered, sealed envelope.'

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Comment: participants nor personnel blinded to treatment allocation. Treatment consisted of the study drug added to standard care versus usual care alone. Usual care was left at the discretion of the treating physician with two participants treated with the study medication versus one participant treated with usual care receiving hypertonic saline. This differential treatment in the two groups could have influenced the outcome in favour of the study drug

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Quote: from ClinicalTrials.gov 'Masking: single blind ‐ outcomes assessor'

Comment: still unclear whether outcome assessors of non patient‐reported outcomes other than Glasgow Coma Scale and NHS Stroke Scale were blinded, but outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Quote: 'all randomised patients completed the protocol'

Comment: no attrition or exclusion

Selective reporting (reporting bias)

Low risk

Comment: protocol registered with ClinicalTrials.gov, reported outcomes match those originally registered

Other bias

Unclear risk

• Bias through possible financial conflict of interest of the authors

  • Quote: 'AMN is responsible for data integrity and the statistical analysis. AMN acknowledges past research support from NovoNordisk, the Neurocritical Care Society, and the Northwestern Memorial Foundation for an unrelated project. AMN acknowledges research support from Gaymar Inc unrelated to this study.'

  • Comment: study commercially sponsored, but declaration of interest available and indicating limited further influence

• Sponsorship bias

  • Quote: 'This study was funded by Astellas Pharma, US. The study sponsor had no role in the design of the protocol, recruitment of patients, analysis or decision to submit for publication. AMN is responsible for data integrity and the statistical analysis. AMN acknowledges past research support from NovoNordisk, the Neurocritical Care Society, and the Northwestern Memorial Foundation for an unrelated project. AMN acknowledges research support from Gaymar Inc unrelated to this study

  • Comment: study commercially sponsored, but declaration of interest available and indicating limited further influence. We felt unclear on what this means for risk of bias

Methods

• Design: parallel RCT (phase not applicable)

• Hyponatraemia subgroup: no

• Recruitment period: 2010 to 2011

• Treatment duration: not reported

• Follow‐up: three months

Participants

• Country: France

• Number of centres: 1

• Inclusion criteria: people admitted to the internal medicine ward or resident of the nursing home of the same institution; ≥ 65 years; serum sodium concentration 123 to ≤ 134 mmol/L; detected on routine biochemistry

• Exclusion criteria: not receiving medications with the ability to cause hyponatraemia; hypervolaemia caused by heart failure or liver cirrhosis; treatment modification not possible according to treating physician; dementia

• Characteristics randomised participants

  • Number: treatment group (9); control group (10)

  • Number, % women: treatment group (5, 56%); control group (6, 60%)

  • Mean age ± SD (years): treatment group (91 ± 4); control group (89 ± 7)

  • Serum sodium concentration (mmol/L): treatment group (132 ± 2); control group (130 ± 2)

  • Cause of hyponatraemia: drug‐induced hyponatraemia (100%)

Interventions

Treatment group

• Change in prescribed medications

Control group

• Standard care

Co‐interventions (all groups)

• Fluid restriction: no restriction

• sodium intake: no restriction save for one patient in treatment group who stopped dietary salt restriction after pharmacologist advice.

Outcomes

• Death

• Response in serum sodium concentration (categorical outcome): serum sodium concentration > 135 mmol/L

• Serum sodium concentration (continuous outcome):change from baseline as well as absolute values

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: not otherwise defined

  • Osmotic demyelination syndrome: not otherwise defined

Notes

• Funding source: 'a pharmacological association'

• Registration number: ID RCB 2010‐A00778‐31

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: 'Les patients ont été randomisés en deux bras...'

Additional information provided by the authors: 'Random sequence was managed as a single randomisation list managed by the sponsor'
Comment: the small number of participants may have caused simple randomisation to fail in producing balanced groups, participants in experimental group had slightly higher baseline serum sodium concentration and took more medications. Consequences for risk of bias judged to be low

Allocation concealment (selection bias)

Low risk

Additional information provided by the authors: 'Allocation was concealed using masking envelope'.

Comment: Allocation concealment method considered adequate

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Additional information provided by the authors: 'No measures were taken to prevent participants or personnel from knowing to which treatment participants had been allocated'

Comment: blinding not attempted and difficult because of intervention type. There were no specific instructions concerning co‐interventions. Differences in co‐interventions may have influenced difference in outcome in favour of experimental treatment

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Additional information provided by the authors: 'No measures were taken to prevent outcome assessors from knowing to which treatment participants had been allocated'

Comment: blinding of outcome assessors not attempted, but main measured outcomes objective.

Incomplete outcome data (attrition bias)
All outcomes

High risk

Comment: overall 26% incomplete outcome data at four weeks; 1 attrition in control group, due to being released from hospital before four weeks; 1 exclusion in experimental group for gastrointestinal bleed after withdrawing proton pump inhibitor; 3 in control group: 2 for stopping a drug possibly causing hyponatraemia ‐ hence possibly biasing results in favour of experimental treatment ‐ 1 due to death

Selective reporting (reporting bias)

Low risk

Comment: protocol available, long‐term study (> 1 week), and both primary outcome death and secondary outcomes related to serum sodium concentration reported. Given lowest serum sodium concentration equalled 129 mmol/L and the intervention consisted of medication withdrawal, potential for rapid increases in serum sodium concentration very limited

Other bias

Low risk

• Bias through possible financial conflict of interest of the authors

• Sponsorship bias

  • Additional information provided by the authors: 'The study was sponsored by a pharmacological association'

  • Comment: Unlikely for commercial sponsorship to be an issue

Methods

• Design: parallel RCT (phase 2a)

• Hyponatraemia subgroup: yes

• Recruitment period: 2007

• Treatment duration: 15 days

• Follow‐up: 45 days

Participants

• Country: Belgium

• Number of centres: 5

• Inclusion criteria: hospitalised participants from university teaching hospitals; 18 to ≥75 years; liver cirrhosis and ascites

• Exclusion criteria: not reported

• Characteristics randomised participants

  • Number: treatment group (10); control group (5)

  • Number, % women: not reported

  • Mean age ± SD (years): not reported

  • Serum sodium concentration (mmol/L): not reported

  • Cause of hyponatraemia: liver cirrhosis (100%)

Interventions

Treatment group

• Oral M0002 (RWJ‐351647): 0.3 mg/d up to 6 mg/d titrated every 3 days

Control group

• Oral placebo

Co‐interventions (all groups)

• Fluid restriction: not reported

• sodium intake: not reported

Outcomes

• Response in serum sodium concentration (categorical outcome): normalisation of serum sodium concentrations = insufficiently reported to allow contribution to meta‐analysis

• Treatment‐specific side‐effects

  • Adverse events: not otherwise defined

  • Serious adverse events: bacterial peritonitis, duodenal perforation and daydreaming

Notes

• Funding source: Early Development, Movetis, Turnhout, Belgium

• Publication: abstract only

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'In a randomised, double‐blind, placebo‐controlled trial, 15...were treated...'

Comment: insufficient information to permit judgement of ‘Low risk’ or ‘High risk’

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of ‘Low risk’ or ‘High risk’

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Quote: 'In a randomised, double‐blind, placebo‐controlled trial, 15...were treated...'

Comment: blinding of participants and personnel attempted. However, lack of detail in reporting of outcomes makes it difficult to assess to what extent it was actually achieved. insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Selective reporting (reporting bias)

High risk

Comment: no protocol, long‐term study (> 1 week) but no reporting of primary outcomes death or health‐related quality of life, no reporting of secondary outcomes related to rapid increase in serum sodium concentration

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Comment: No declaration of interest. Last two authors' affiliations are linked to the commercial company developing the medication

• Sponsorship bias

  • Quote: 'Early Development, Movetis, Turnhout, Belgium'

Methods

• Design: parallel RCT (phase 2)

• Hyponatraemia subgroup: no

• Recruitment period: 2008 to 2009

• Treatment duration: 7 days

• Follow‐up: 14 days

Participants

• Country: China

• Number of centres: 9

• Inclusion criteria: 18 to 75 years; serum sodium concentration < 135 mmol/L; SIADH

• Exclusion criteria: serum sodium concentration < 120 mmol/L with apathy, clouded consciousness and convulsion; hypovolaemia; acute of transient hyponatraemia cause by brain trauma or surgery; uncontrolled hypothyroidism or adrenal insufficiency; cardiac surgery, sustained ventricular tachycardia or ventricular fibrillation and not using cardioverter defibrillator; severe stable or unstable angina, cerebrovascular accident within 30 days of recruitment; psychogenic polydipsia; systolic BP < 90 mm Hg; history of hypersensitivity of benzodiazepines; drug abuse; fasting glucose concentration > 220 mg/dL; advanced terminal illness or severe pulmonary hypertension; drug‐induced hyponatraemia excluding diuretics; creatinine > 1.5 x upper limit; central nervous system diseases; transaminase > 2.5 x upper limit; requiring intravenous fluid for any reason; use of vasopressin analogues; treatment of hyponatraemia within 13 days of recruitment; nursing or pregnancy

• Characteristics randomised participants

  • Number: treatment group (21); control group (24)

  • Number, % women: treatment group (11, 52%); control group (9, 38%)

  • Mean age ± SD (years): treatment group (63 ± 14); control group (61 ± 13)

  • Serum sodium concentration (mmol/L): treatment group (127 ± 5); control group (125 ± 6)

  • Cause of hyponatraemia: SIADH (100%)

Interventions

Treatment group

• Oral tolvaptan: 1 x 15 mg/d up to 1 x 60 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: 'fluid restriction (1500 mL/d) was added at the discretion of the investigator when the patient's serum sodium levels were below 130 mmol/L persistently'

• sodium intake: not reported

Outcomes

• Death

• Response: serum sodium concentration ≥ 135 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline

• Treatment‐specific side‐effects

  • Thirst: not otherwise defined; insufficiently reported to allow contribution to meta‐analysis

Notes

• Funding source: Otsuka Pharmaceutical

• Trial acronym: We allocated the trial acronym 'Otsuka study' from the quote 'the present study analysed SIADH patients recruited in the ....Otsuka study', to indicate the three separately reported subgroups belonged to the same trial

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: 'This randomised double‐blind placebo‐controlled trial...' and 'Randomisation was stratified by underlying disease and whether the hyponatraemia was mild...or marked...'

Comment: since the investigators went to the trouble to stratify allocation, it seems probable enough that sequence generation was adequate

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Quote: 'This randomised double‐blind placebo‐controlled trial...' and 'Patients received tolvaptan or matching placebo..'

Comment: blinding probably attempted, although due to increase in urine output, un‐blinding likely to have occurred. Fluid restriction was largely left to the discretion of the investigator, but fewer participants receiving the study medication had fluid restriction initiated or tightened by day 30 (11% versus 21%)

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: No attrition for main outcomes

Selective reporting (reporting bias)

High risk

Comment: the study was registered with ClinicalTrials.gov, but only primary efficacy outcome clearly listed, short‐term study (≤1 week) death reported, but no reporting of secondary outcomes related to rapid increase in serum sodium concentration. Also study results ‐ incompletely/differentially ‐ reported for the subgroups of SIADH, heart failure and cirrhosis separately

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Comment: Insufficient information to permit judgement of 'High risk' or 'Low risk'

• Sponsorship bias

  • Quote: 'This study was funded by Otsuka Pharmaceutical, which provided consultancy fees or honorarium, support for travel to meetings and for language polishing.'

Methods

• Design: parallel RCT (phase 2)

• Hyponatraemia subgroup: no

• Recruitment period: 2008 to 2009

• Treatment duration: 7 days

• Follow‐up: not reported

Participants

• Country: China

• Number of centres: 11

• Inclusion criteria: inpatients, hospitalised for congestive heart failure; 18 to 80 years; serum sodium concentration < 135 mmol/L

• Exclusion criteria: hypovolaemia; haemodynamic instability; insertion of heart valves; open heart surgery or artificial pacemaker insertion within 60 days of recruitment; heart valve insertion or heart transplant within 30 days of recruitment; acute MI; cardiac hypertrophy; active myocarditis or amyloid cardiomyopathy; severe aortic stenosis; severe pulmonary hypertension; uncontrolled low function of thyroid or adrenal glands; epilepsy; Guillain‐Barre Syndrome; anuria; ventricular tachycardia; ventricular fibrillation; implantable cardioverter defibrillators; cerebral vascular accidents within 30 days of recruitment; allergy to benzodiazepines; medication or alcohol abuse; BMI > 35 kg/m; supine systolic pressure < 90 mm Hg; abnormal laboratory results at the time of recruitment: fasting blood sugar > 12.2 mmol/L, haemoglobin <80 g/L, total bilirubin > 85.5 µmol/L, SCr > 2 x the upper limit, potassium > 5.5 mmol/L, transaminases > 5 x the upper limit, sodium < 120 mmol/L

• Characteristics randomised participants

  • Number: treatment group (35); control group (30)

  • Number, % women: treatment group (15, 43%); control group (12, 40%)

  • Mean age ± SD (years): treatment group (63 ± 15); control group (65 ± 15)

  • Serum sodium concentration (mmol/L): treatment group (130 ± 4); control group (131 ± 3)

  • Cause of hyponatraemia: heart failure (100%)

Interventions

Treatment group

• Oral tolvaptan: 1 x 15 mg/d up to 1 x 60 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: not reported

• sodium intake: not reported

Outcomes

• Death

• Serum sodium concentration (continuous outcome): change from baseline

• Outcomes related to over‐correction of serum sodium concentration

  • Hypernatraemia: not otherwise defined

• Treatment‐specific side‐effects

  • Hypotension: systolic BP < 100 mm Hg or new vasopressor medication

Notes

• Funding source: Otsuka Pharmaceutical

• Primary paper in Mandarin ‐ Translated by Sunny Wu

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: 'This randomised double‐blind placebo‐controlled trial...' and 'Randomisation was stratified by underlying disease and whether the hyponatraemia was mild...or marked...'

Comment: since the investigators went to the trouble to stratify allocation, it seems probable enough that sequence generation was adequate

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Quote: 'This randomised double‐blind placebo‐controlled trial...'

Comment: blinding of participants and personnel attempted. However, lack of detail in reporting of outcomes makes it difficult to assess to what extent it was actually achieved. insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Comment: No attrition for main outcomes

Selective reporting (reporting bias)

High risk

Comment: the study was registered with ClinicalTrials.gov, but only primary efficacy outcome clearly listed, short‐term study (≤ 1 week) but no reporting of primary outcomes death or health‐related quality of life, no reporting of secondary outcomes related to rapid increase in serum sodium concentration in a way that permitted data extraction. Also study results ‐ incompletely/differentially ‐ reported for the subgroups of SIADH, heart failure and cirrhosis separately

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • Comment: Insufficient information to permit judgement of 'High risk' or 'Low risk'

• Sponsorship bias

  • Quote: 'The study was conducted by Otsuka Beijing Research Institute. The institute is funded by Otsuka Pharmaceutical CO., Ltd. Japan.'

Methods

• Design: parallel RCT (phase not reported)

• Hyponatraemia subgroup: no

• Recruitment period: not reported

• Treatment duration: 7 days

• Follow‐up: not reported

Participants

• Country: China

• Number of centres: not reported

• Inclusion criteria: liver cirrhosis

• Exclusion criteria: not reported

• Characteristics randomised participants

  • Number: not reported

  • Number, % women: not reported

  • Mean age ± SD (years): not reported

  • Serum sodium concentration (mmol/L): not reported

  • Cause of hyponatraemia: liver cirrhosis (100%)

Interventions

Treatment group

• Oral tolvaptan: 1 x 30 mg/d up to 1 x 60 mg/d

Control group

• Oral placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: not reported

• Dietary sodium intake: not reported

Outcomes

• Response: normalisation of serum sodium concentration; insufficiently reported to allow contribution to meta‐analysis

• Serum sodium concentration (continuous outcome) change from baseline until day 4 and 7; insufficiently reported to allow contribution to meta‐analysis

Notes

• Funding source: Otsuka Pharmaceutical

• Publication: abstract only

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: 'A multicenter, randomised, double‐blind,..', and 'Patients were randomised to receive...'

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Allocation concealment (selection bias)

Unclear risk

Comment: insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Quote: 'A multicenter,..., double‐blind, placebo‐controlled clinical trial...', and 'Patients were randomised to receive placebo or tolvaptan...'

Comment: 'insufficient information to permit judgement of 'High risk' or 'Low risk'

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Comment: unclear whether assessors of non patient‐reported outcomes were blinded, but main measured outcomes objective

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Comment: 'insufficient information to permit judgement of 'High risk' or 'Low risk'

Selective reporting (reporting bias)

High risk

Comment: the study was registered with ClinicalTrials.gov, but only primary efficacy outcome clearly listed,, short‐term study (≤ 1 week) but no reporting of primary outcomes death or health‐related quality of life, and no reporting of secondary outcomes related to rapid increase in serum sodium concentration in a way that permitted data extraction. Also study results ‐ incompletely/differentially ‐ reported for the subgroups of SIADH, heart failure and cirrhosis separately

Other bias

High risk

• Bias through possible financial conflict of interest of the authors

  • No declaration of interest

• Sponsorship bias

  • Quote: 'This study was funded by Otsuka Pharmaceutical, which provided consultancy fees or honorarium, support for travel to meetings and for language polishing.'

Methods

• Design: parallel RCT (phase 3)

• Hyponatraemia subgroup: no

• Recruitment period: 2011 to 2012

• Treatment duration: 14 days

• Follow‐up: 30 days

Participants

• Country: USA

• Number of centres: 1

• Inclusion criteria: patients admitted to University of Texas MD Anderson cancer centre; ≥ 18 years; serum sodium concentration 125 to ≤ 130 mmol/L; non‐hypovolaemia; Eastern Cooperative Oncology Group performance status from 0 (fully active) to 3 (capable of only limited self‐care and confined to bed or chair ≥50% of waking hours)

• Exclusion criteria: critical illness; GFR < 25 mL/min; correctable hyponatraemia; use of demeclocycline, lithium or CYP 3A4 modulators; life‐expectancy < 3 months

• Characteristics randomised participants

  • Number: treatment group (24); control group (24)

  • % women: treatment group (47%); control group (46%)

  • Median age, range (years): treatment group (69, 52 to 81); control group (60, 20 to 85)

  • Serum sodium concentration (mmol/L): treatment group (128 ± 2); control group (129 ± 1)

  • Cause of hyponatraemia: cancer (100%)

Interventions

Treatment group

• Oral tolvaptan: 1 x 15 mg/d up to 60 mg/d

Control group

• Oral matching placebo: 1 x 1/d

Co‐interventions (all groups)

• Fluid restriction: fluid restriction < 1.5 L/d, but participants were allowed to drink to thirst

• Dietary sodium intake: not reported

Outcomes

• Death

• Length of hospital stay (continuous outcome): insufficiently reported to allow contribution to meta‐analysis

• Cognitive function (Mini Mental State Exam score (range 1 to 16)): change from baseline

• Response: serum sodium concentration ≥ 136 mmol/L

• Serum sodium concentration (continuous outcome): change from baseline

• Treatment‐specific side‐effects

  • Hypotension: not otherwise defined

  • Thirst: not otherwise defined

  • Polyuria: increase urination frequency and volume

  • Serious adverse events: not otherwise defined

• Treatment discontinuation: insufficiently reported to allow contribution to meta‐analysis

Notes

• Funding source: Otsuka America Pharmaceuticals Inc.