Clarifying Optimal Sodium InTake In Cardiovasular and Kidney (COSTICK) Diseases: a study protocol for two randomised controlled trials

Study aim and design

Both trials were designed to explore the effects of an educational intervention to reduce sodium intake targeting low levels, compared to usual care. The Sodium In Take In Chronic Kidney Disease (STICK) clinical trial was designed to explore the effect of the intervention in adults with non-severe CKD on decline in renal function over two years. The Clarifying Optimal Sodium Intake in Populations (COSIP) clinical trial was designed to explore the effect of the intervention in adults without cardiovascular disease on a panel of cardiovascular biomarkers, over two years.

Both trials are Phase IIb randomised, two-group, parallel, open-label, controlled trials conducted at a single-centre. Blinding of participants and the research dietitian was not possible for practical reasons. Participants were recruited from hospital-based outpatient clinics or attending general practitioners in the catchment area of a tertiary referral hospital in Galway, Ireland. Potentially eligible participants were invited to participate and provided written informed consent before entering a screening period, consisting of an abbreviated food frequency questionnaire (FFQ; see extended data²⁴), medical history, laboratory investigations, baseline physical measurements (height, weight, body mass index (BMI), hip and waist circumference) and blood pressure (BP). Office BP was measured using a calibrated, automated oscillometric device and followed by a 24-hour ambulatory blood pressure monitor (ABPM) using the Spacelabs ABP 90217 device, where BP was measured every 30 minutes between 7am and 10pm, and every 60 minutes between 10pm and 7am. Given the complementary nature of both trials, participants were screened for both trials simultaneously, but only deemed eligible and randomized into STICK or COSIP. Rescreening of participants who failed the screening visit due to a blood pressure result or medication change was permitted after a minimum of three months.

Eligibility criteria for both trials

Inclusion criteria for both trials included age >40 years, stable blood pressure, unchanged anti-hypertensive medications, willingness to modify dietary intake and written informed consent (Table 1). Initially, we also required dietary assessed sodium intake of >2.3g/day, but removed this after the Vanguard phase due to weak correlation between the dietary sodium assessment and baseline 24 hour urine sodium collection. Exclusion criteria for both trials included abnormal sodium handling, heart failure, high dose diuretic use, immunosuppressive medication use, unable to comply with intervention or study visits, pregnancy or lactation, postural hypotension, cognitive impairment, high or low BMI or inclusion in another clinical trial.

Eligibility criteria – STICK specific

All STICK participants were required to have a stable estimated glomerular filtration rate (eGFR) of 30–60ml/min/1.73m² within three months of randomization. Participants were excluded for recent acute kidney Injury (AKI), rapidly declining eGFR²⁵, known glomerular disease or current use of non-steroidal anti-inflammatory drugs (NSAID).

Eligibility criteria – COSIP specific

For COSIP, participants were excluded for known CKD or CVD.

Randomisation

Once confirmed eligible, participant details were entered into a trial-specific, secure, web-based randomisation service. In STICK, we used a minimization strategy (including previous attendance at a dietitian, diabetes, hypertension and strata of eGFR [30–39, 40–49 and 50–60ml/min/1.73m²]) to support computer-generated block randomization with 1:1 assignment to intervention or usual care. In COSIP, we used computer-generated block randomization to 1:1 assign participants to intervention or usual care. Variable block sizes were used to minimize the occurrence of chance imbalances and to preserve allocation concealment. After confirmation of eligibility, participants may have been randomised in advance to facilitate the scheduling of individuals to attend distinct clinics for either the intervention or usual care arms.

Study interventions

Usual care: All participants received a one-to-one, dietitian-developed healthy eating guidance session²⁶, administered by a trained member of the research team, over 15 minutes following randomization, in addition to written materials emphasizing key messages²⁷. Telephone contact was made at months nine and 15 to follow-up on key points on healthy eating (each 15 minutes) with a total contact time of 45 minutes over the course of the trials. Participants randomised to usual care did not meet with the dietitian nor did they receive focused recommendations on sodium intake (other than general guidance contained in 'Health Food for Life'²⁷).

Sodium reduction intervention: In addition to healthy eating guidance sessions, those randomised to the intervention received specific counseling on behavioural and environmental factors to promote reduction in sodium intake to a target of <100mmol/day (<2.3g/day). A research dietitian developed the specific components of the intervention, based on the trial of non-pharmacologic interventions in the elderly (TONE) trial²⁸. A registered dietitian delivered the intervention using an operations manual to standardize delivery within and across participants. In circumstances where the research dietitian was unavailable, an appropriately trained delegate delivered the intervention. The intervention targeted: (i) reducing use of ‘salt’ during food preparation (encouraging the use of herbs and/or spices); (ii) reducing table ‘salt’ use; (iii) encouraging fresh food consumption over processed or canned foods; (iv) identification of sodium content in foods; (v) modifying the consumption of foods with high sodium content; and (vi) advice on eating outside of home. After determining health literacy, using the Newest Vital Sign UK (NVS-UK)²⁹, participants were counselled and behaviour change techniques (goal setting, self-monitoring, frequent contact, feedback and re-enforcement, problem solving and motivational interviewing) were employed to reduce sodium intake by setting specific, measurable, achievable, realistic and timely (SMART) goals for the individual. Tools used included 24-hour dietary recall (analysed using Nutritics³⁰), three-day food diaries and sodium behavior checklists.

The intervention was delivered at all patient contacts up to and including T₇ (21 month) visit (Table 2) with approximately 225–255 minutes of in-person contact and 155 minutes of telephone contact over the course of the trials. Participants had in-person, one-to-one sessions with the study dietitian at randomization (T₀), vanguard (T_V), where relevant, and T₂; group sessions at T₁ and T₄. In addition, each participant received five separate 15 min telephone calls (Week 1, Week 2, Week 4, Week 6 and Week 10) and 20 min telephone calls at T₃, T₅, T₆ and T₇.

Follow-up schedule

To minimize participant inconvenience and burden, six or seven in-person visits were considered essential (screening, randomization (T₀), T_v (first 20 participants for each trial), T₁, T₂, T₄ and T₈) and the remaining four follow-up visits were carried out by telephone (T₃, T₅, T₆ and T₇) (Table 3). Post-trial, participants will return to routine clinical care.

Data collection

The following data were collected using standardised case report forms (CRF): demographics, physical measurements (BP, height, weight, waist and hip circumference), medications, medical history, lifestyle factors, randomization allocation, diet, delivery of healthy eating guidance session, delivery of the sodium reduction intervention, clinical events, hospitalisations, adverse events and laboratory results (see extended data²⁴). Diet was measured using FFQ, food diaries and a questionnaire focused on sodium-specific behaviours (see extended data²⁴). Blood samples were drawn at the time of clinic attendance from participants in the seated position and using a tourniquet. For renin and aldosterone, an EDTA sample was drawn and spun at 2000g for 10minutes at 20 degrees Celsius within 30minutes of phlebotomy. Plasma was then frozen at -20 or -80 degrees Celsius until periodically transferred to our local laboratory using styrofoam boxes and cooler packs to ensure samples did not defrost. Laboratory samples were analysed at the Galway University Hospital laboratory using standardised storage, handling and analytical procedures. We measured serum sodium, potassium, creatinine, glycated haemoglobin (HbA1c), high-density lipoprotein (HDL), low-density lipoprotein (LDL), direct renin, aldosterone, high sensitivity troponin T (hsTnT), Pro B-type Natriuretic Peptide (proBNP), C-reactive protein (CRP), urine protein creatinine ratio (uPCR), 24 hour urine (sodium, potassium, creatinine and protein) and fasting morning urine (sodium, creatinine and protein). This included serum creatinine, measured using methodology traceable to an isotope dilution mass spectrometry (IDMS) reference measurement. Samples were transported to the laboratory within one hour of blood draw and analyses were conducted using standardized laboratory methods on the Roche Cobas® 8000 modular analyser series (Roche Diagnostics Limited, West Sussex, UK), except for HbA1c which was measured using capillary electrophoresis on the Sebia Capillarys 3 automated haemoglobin analyser using the Sebia HbA1c kit (2017/05). After additional consent was obtained, additional blood and urine samples, as well as PAXgene samples for RNA measurement, for biobanking were taken at T₀, T₁ and T₈. To ensure high quality data collection, staff received training on completing paper-based CRF and additional training on entering data into the study database, which included database integrity checks to ensure data consistency.

Study outcomes

In STICK, the primary outcome measure is change in 24-hour urinary creatinine clearance over two years. We adopt this approach, rather than eGFR, to improve reliability and validity of the primary outcome measure. In COSIP, the primary outcome measure is change in cardiovascular biomarkers (renin, aldosterone, hsTnT, proBNP and CRP) over two years. For COSTICK, the primary outcome measures are creatinine clearance, renin, aldosterone, hsTnT, proBNP and CRP. Secondary outcomes included change in (i) 24 hour ABPM, (ii) 24 hour urine sodium excretion, (iii) 24 hour urine protein, (iv) Modification of Diet in Renal Disease (MDRD) eGFR²⁵, (v) Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) eGFR³¹, (vi) Kidney Disease Improving Global Outcomes (KDIGO) CKD risk category³² and (vii) change in PAXgene RNA, and number of (viii) cardiovascular events (acute coronary syndrome, myocardial infarction, heart failure or stroke), (ix) syncope events (prescyncope, syncope, hypotension or orthostatic hypotension) and (x) renal events (acute kidney injury, dialysis, transplant or end stage renal disease).

Safety reporting

An adverse event is defined as any untoward medical occurrence, including an exacerbation of a pre-existing condition or abnormal laboratory finding, with or without a causal relationship with sodium intake. Participants were monitored for adverse events from informed consent through completion of the final study visit. All adverse events were captured on case report forms, entered into the clinical trial database and coded using the Medical Dictionary for Regulated Activities (MedDRA). Annual line listings of adverse events were reported to the ethics committee.

Withdrawal of intervention

The study intervention was withdrawn if: (i) symptomatic orthostatic hypotension with systolic BP ≤100mmHg; (ii) diagnosis of glomerular disease (as listed in exclusion criteria); (iii) diagnosis of ESRD, initiation of dialysis or renal transplantation; (iv) prescription of sodium bicarbonate therapy or treating physician recommended against sodium restriction; (v) pregnancy; or (vi) acquired new dietary requirement that precluded the intervention. These participants were encouraged to complete all remaining follow-up visits.

Sample size calculation

For STICK, we assumed a mean decline in creatinine clearance of 8±5ml/min/1.73m² over the trial period³³ and estimated a minimum clinically meaningful effect size of 25% relative reduction. Based on an alpha of 0.05 and power of 80%, a per group sample size of 99 participants was required. Assuming a dropout rate of 5% participants, a net crossover/non-adherence of 5% in favour of the control group, we required a total sample size of 224 participants. For COSIP, we based sample size on the ability to detect an effect size of 0.40 in the between-group difference in mean change scores of biomarkers (80% power and alpha 0.05), equating to a difference of 0.4 of the standard deviation of the change in biomarker. As the COSIP trial is considered exploratory, each biomarker is tested at significance threshold p<0.05. Assuming a dropout rate of 5% participants, a net crossover/non-adherence of 15%, we require a total sample size of 286 participants. Combining COSIP and STICK trials resulted in the ability to detect a smaller effect size.

Statistical analysis plan

Descriptive statistics will be used to describe baseline characteristics, the flow of trial participants and the amount of missing data for predictor and outcome variables. All losses to follow-up and dropouts will be accounted for and reasons documented. Participants randomised but who did not attend the randomisation visit will be included in analyses. Suitable transformations will be used if the underlying normality assumption is a concern and the bootstrap procedure and a comparison of the median change will be used if a suitable transformation cannot be made. All tests of significance will be two-sided using an alpha of 0.05 for statistical significance and will not be adjusted for multiple testing due to the exploratory nature of the trials. The primary analysis will use the intention-to-treat dataset for all outcomes. A secondary per-protocol analysis will also be carried out, with post randomisation exclusions due to ineligibility, non-compliance, loss to follow-up or missing data.

The primary analysis will consist of fitting a separate linear, multivariable analysis to each of the six outcomes: renin, aldosterone, creatinine clearance, pro-BNP, troponin, and CRP, and secondary outcome measure of 24-hour ABPM. We will review data for implausible values and exclude extreme changes (e.g. change in renin or aldosterone impacted by change in medications). Corresponding baseline measurements of each biomarker will be included as covariates in the model, along with a factor variable for the sub-study in which each patient was recruited, age, sex and whether the participant had attended a dietitian in the 6 months previous to baseline (incident vs prevalent participant). To ensure estimates are more robust to potential effects of missing data at 24-month follow-up, an inverse probability weighting will be applied in all regression analyses to account for the differing propensity of patients with certain baseline characteristics to be included in the final analysis. All covariates included in the regression model will be included in the missingness model, including the other 5 primary endpoints.

Reporting of primary analysis results will consist of tabulations of the unadjusted and adjusted mean differences at 24 months between intervention and control groups for each biomarker, with confidence intervals of these parameter estimates and p-values, along with descriptive unadjusted mean differences at 24 months between intervention and control groups. A forest plot of the standardised, adjusted mean difference for all 6 endpoints will also be generated. Statistical tests of interaction (Wald) will be used for all subgroup analyses. A sensitivity analysis will be performed to assess the effect of participant reported compliance on treatment effect. Due to the duration of this trial, we expect to have a complete dataset for covariate and outcome data for 90% of participants. Complete case analysis will be used to handle missing data. A sensitivity analysis with multiple imputation will also be carried out if there is >10% missing data for covariates.

Data monitoring

A data and safety monitoring board (DSMB), chaired by a clinician scientist with experience in clinical trials, reviewed the progress of the trials and safety data when 50% of participants completed one-year follow-up, when the DSMB recommended that the trials proceed. An independent study monitor reviewed the integrity of study data at multiple timepoints throughout the trials and ensured that study procedures were compliant with the International Conference on Harmonisation Good Clinical Practice (ICH-GCP).

Study administration

These trials are single centre studies with all activities conducted and coordinated at the HRB Clinical Research Facility Galway (HRB-CRFG). The investigators did not implement any deviation from, or changes to, the trial protocols without review and documented approval from the research ethics committee. All participants provided written informed consent before any study specific activities were completed, and all data was handled with strict privacy and electronic data security standards, using unique subject identifiers to prevent unauthorised identification of research participants. Pseudononymised data were entered into a study-specific, secure, password-protected clinical trial database with an audit trail.

Dissemination of information

The results of the trials will be presented at scientific meetings (local, national and international), patient-oriented meetings and published in peer-reviewed publications. The HRB-CRFG and investigators have the ownership of all data and results collected during the trials and have full rights to publication, without restriction. For all publications and presentations from this work, authors will be required to make substantial contributions to (i) concept or design; acquisition, analysis or interpretation of data; (ii) drafting or critically revising manuscripts and (iii) give final approval of the version to be published, in line with the International Committee of Medical Journal Editors guidelines for authorship. The datasets will not be made publicly available, instead they will be maintained at HRB-CRFG and the investigators must review and approve all requests for access to the data.

Study status

Recruitment of both clinical trials is closed and the final visits of all patients have taken place. Data cleaning is ongoing, ahead of database lock, statistical analyses and trial reporting.

Underlying data

No data are associated with article

Extended data

Open Science Framework: Clarifying Optimal Sodium InTake In Cardiovasular and Kidney (COSTICK). https://doi.org/10.17605/OSF.IO/HYKT5²⁴

This project contains the following extended data:

Reporting guidelines

Open Science Framework: SPIRIT Checklist for “Clarifying Optimal Sodium In Take In Cardiovasular and Kidney (COSTICK) Diseases: a study protocol for two randomised controlled trials”, https://doi.org/10.17605/OSF.IO/HYKT5²⁴

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain deciation).