Intensification of blood pressure lowering therapeutics based on diuretics versus usual management for uncontrolled hypertension in patients with moderate to severe chronic kidney disease (THINK): protocol for an open label, two-parallel group, cluster randomised controlled, superiority, phase 3 trial Authors: Floriane Le Vilain-Abraham, Jean-Michel Halimi, Patrick Rossignol, Maurice Laville, Michel Azizi, Bénédicte Stengel, Valentin Maisons, Agnès Rouzès, Agnès Caille, Bénédicte Sautenet Published: 2026 ABSTRACT AbstractIntroductionHypertension is the primary modifiable risk factor for cardiorenal complications of chronic kidney disease (CKD). It is well established that lowering blood pressure (BP) in patients with CKD reduces renal and cardiovascular complications. The guidelines advise starting the BP-lowering drug with ACE inhibitors (ACEi) or angiotensin receptor blockers (ARB), but then there is no strong evidence to support the preferential use of any particular agent in controlling BP. Our hypothesis is that patients with CKD and uncontrolled hypertension are fluid overloaded and that the second line of treatment after an ACEi or an ARB should be a diuretic.Methods and analysisThis is a two-parallel group cluster randomised superiority phase 3 trial comparing a specific algorithm to lower BP in patients with moderate to severe CKD based on diuretics to standard of care. The primary endpoint is a composite time-to-event endpoint of (1) kidney failure with replacement therapy (KFRT), (2) estimated glomerular filtration rate decline ≥40%, (3) cardiovascular events (myocardial infarction, heart failure hospitalisation, stroke) and (4) all-cause mortality. Secondary endpoints include each component of the primary endpoint, systolic and diastolic BP at 3 months and 6 months post baseline then every 6 months, proportion of patients with controlled BP at 2 years, change from baseline in glomerular filtration rate and in proteinuria (g/day) or proteinuria/creatininuria (g/g) at 3 months and 6 months post baseline then every 6 months, proportion of patients who used at least one diuretic, quality of life and safety. It is planned to include 720 patients in 40 centres.Ethics and disseminationThis study was approved by the Ethics Committee CPP Ile-de-France VII on 23 January 2023 (ref: 2022–5 01 494-39-00). Study findings will be published in a peer-reviewed journal and presented at scientific conferences.Trial registration numberNCT05732727. STRENGTHS AND LIMITATIONS OF THIS STUDY To prevent group contamination, cluster randomisation was chosen and the algorithm is not disclosed in the protocol. To prevent group contamination, cluster randomisation was chosen and the algorithm is not disclosed in the protocol. Blood pressure will be assessed at home by patients with a device furnished by the study. Blood pressure will be assessed at home by patients with a device furnished by the study. The very nature of the intervention assessed prevents any form of blinding, neither for care providers nor for patients. The very nature of the intervention assessed prevents any form of blinding, neither for care providers nor for patients. To limit the bias due to the unblinded design, the primary outcome will be assessed by a blinded adjudication committee. To limit the bias due to the unblinded design, the primary outcome will be assessed by a blinded adjudication committee. Introduction Background and rationale All over the world, 1.2 million people die from chronic kidney disease (CKD) every year and this is projected to reach at least 2.2 million by 2040.1 The burden of CKD on public health is major and CKD is the 12th leading cause of death worldwide.2 Patients with CKD are up to 10 times more likely to die prematurely than progressing to kidney failure requiring kidney replacement therapy, largely due to cardiovascular disease.3–5 CKD is defined as a reduced kidney function (estimated glomerular filtration rate (eGFR) under 60 mL/min/1.73 m² or kidney damage for 3 months or more). Two large general population cohorts (n=1 464 767) described a loss of life expectancy due to cardiovascular disease by 7 years and 12.5 years in patients with eGFR stages 3b and 4 (30–44 mL/min/1.73 m² and 15–29 mL/min/1.73 m², respectively).6 The burden of symptoms and impaired quality of life in CKD may be compounded by comorbidities, mostly due to cardiovascular diseases.7 8 The first risk factor associated with cardiovascular events in patients with CKD is hypertension. Hypertension is both a cause and effect of CKD and its prevalence in patients with moderate to advanced CKD ranges from 80% to 100%.9–12 Uncontrolled hypertension is a risk factor for developing CKD and progression of CKD. CKD can promote uncontrolled hypertension due to volume expansion and an increase in systemic vascular resistance. Furthermore, there is a joint contribution of CKD and hypertension to cardiac risk.13 There is strong evidence for lowering blood pressure (BP) in patients with CKD in order to slow the progression of renal disease and reduce cardiovascular morbidity and global mortality.14 Even though the drug therapy strategy is well described in the general population (renin-angiotensin-aldosterone system (RAAS) inhibitors (ACE inhibitors (ACEi) and angiotensin receptor blockers (ARB)), diuretics (thiazides and thiazide-like diuretics) and calcium channel blockers (CCB) then mineralocorticoids antagonists in addition for resistant hypertension), there is no strong consensus in nephrology for patients with CKD.15 The guidelines advise starting with ACEi or ARB. However, there is no formal evidence to support the preferential use of any particular agent in lowering BP in CKD patients and the results of clinical trials are discordant.16 There are no randomised trials in patients with CKD comparing different drug combinations or algorithms. Furthermore, the adaptation of hypertensive therapy becomes more complex in patients with stage 4 CKD, with few trials assessing these patients. The international Network of Chronic Kidney Disease (iNET-CKD) reported in 2019 the prevalence of uncontrolled BP in adults with CKD and the heterogeneity of the use for antihypertensive therapy across 17 cohorts (34 602 participants).11 In this study, the highest mean systolic BP was described in the CKD-REIN French cohort (n=2147 participants), compared with the other international cohorts, and the prevalence of uncontrolled BP was 84%. In the CKD-REIN French cohort, among the 91% participants with hypertension, 81.6 % patients had a RAAS inhibitor prescribed, 54.3% a CCB and only 58.2% a diuretic. Our research question concerns the improvement of key clinical events with BP lowering management in CKD patients with moderate and advanced stages 3b and 4. Most patients with CKD need a combination of several antihypertensive drug classes to obtain optimum BP control. In the iNET-CKD study, 49.8% of participants from the CKD-REIN French cohort had three or more antihypertensive drug classes and in the CKD-REIN French cohort, only 9% of participants had the association of three drugs recommended (RAAS inhibitor, CCB and diuretic).11 17 Even with three or more antihypertensive drug classes resistant, hypertension remains high with a prevalence of 40.4% in the CRIC study, a cohort of 3330 patients with CKD.18 As reported above, RAAS inhibitors are recommended as first-line treatment of hypertension in patients with CKD. Several trials studied ACEi and ARB in diabetic kidney disease and hypertensive kidney disease with long follow-up and improvement in clinical outcomes: cardiovascular events, kidney failure with replacement therapy (KFRT) and death.19–23 At that time, the best sequence of drug classes to combine was unknown. The ACCOMPLISH trial studied the effect of benazepril-amlodipine (ACEi-CCB) versus benazepril-hydrochlorothiazide (ACEi-thiazide) in 11 506 patients with hypertension,24 not in CKD population. The combination of ACEi-CCB was superior in reducing kidney progression and cardiovascular events but few patients had a moderate CKD without any patients with advanced CKD. In patients with CKD, the combination of ACEi-CCB was associated with more peripheral oedema and angio-oedema.25 Diuretics are useful therapeutics in patients with CKD for lowering BP and managing swelling.26–28 In association with ACEis, some trials describe a reduction in cardiovascular risk.29 30 However, the other BP lowering drugs (beta-blockers, alpha-blockers or sympatholytic agents) did not show obvious reduction in cardiovascular and kidney failure. Lastly, mineralocorticoid receptor antagonists are not yet recommended in advanced CKD because of the high risk of hyperkalaemia and acute kidney injury. There is a rationale for the use of diuretic agents in CKD patients with hypertension. Kidney dysfunction interferes with salt excretion, leading to volume overload and peripheral vascular resistance rise. The addition of diuretics to other hypertensive drug classes is known to potentiate the BP lowering effects. There is strong evidence for a major implication of volume overload in CKD patients with hypertension. In the NephroTest cohort, a French cohort of 2015 patients with CKD stage 1 to 5, 1782 had hypertension, only 54% had a diuretic and 44% had uncontrolled hypertension.31 In this cohort, extracellular fluid (ECF) overload was an independent determinant of hypertension, uncontrolled hypertension and apparent treatment resistant hypertension. In the same cohort, ECF overload was independently associated with KFRT and death.32 ECF increased by 136 mL per year on average even so diuretic prescription remained stable and the trajectory of ECF was associated with KFRT (adjusted HR (aHR): 1.12 (95% CI 1.06 to 1.18); p<0.001 per 1 L increase in ECF) and death (aHR: 1.10 (95% CI 1.04 to 1.17); p=0.002).33 The diuretics are the preferred drugs in ECF overload situation and there is a rationale for use of different diuretics acting on the three segments of renal tubule in order to reduce BP and impact on diastolic dysfunction and arterial stiffness.34 35 In front of these previous results and the patient clinical profile, the secondary effects attributed to diuretics related to depletion will be minimal. Hypothesis and objectives We hypothesised that a specific strategy to lower BP in patients with moderate to severe CKD based on diuretics combined with an ACEi or an ARB will be more effective on cardiovascular events, KFRT and death compared with standard of care. The primary objective of the study is to assess the efficacy of a BP lowering algorithm based on diuretic therapeutics in reducing the incidence of KFRT, CKD progression, cardiovascular events and mortality in patients with stages 3b and 4 CKD. The secondary objectives are to assess each component of the primary objective separately and then to assess the impact of this algorithm on systolic and diastolic BP, hypertension control, kidney function, proteinuria, use of diuretics, quality of life and safety. Methods and analysis Trial design and setting The trial is planned as an open label, two-parallel group cluster randomised, superiority, phase 3 trial. Cluster randomisation is mandatory in this context because the intervention assessed applies at the physician level. Indeed, it is the physicians (in the intervention group) who will apply the algorithm. Cluster randomisation therefore aims at preventing group contamination. The clusters are French hospitals with nephrology departments. We may have randomised physicians rather than centres, but this would have supposed an absence of communication among physicians working within a same centre, which is unrealistic. We therefore decided to randomise centres. Centres have given their consent to participate based on inclusion and exclusion criteria of the study and they completed a feasibility survey. The trial involves 40 French nephrology centres from university and peripheral hospitals (listed in the online supplemental table 1). The timeline of the trial36 is displayed in the figure 1. Timeline cluster diagram of the THINK trial. Participants Eligible participants will be male or female over 18 years old, with a Clinical Frailty Score ≤5 for patients aged over 80 years, and with an advanced or moderate CKD (eGFR 15.0–44.9 mL/min/1.73 m² using Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula 2021). Participants should have arterial hypertension treated with at least one BP lowering drug therapy among blockers of the renin-angiotensin system (ACEi or ARB), at the maximal posology tolerated by the patients, stable since at least one month. Other BP lowering drug therapies are tolerated without limit. Patients with only one diuretic including thiazide or loop diuretic could be included. Uncontrolled office BP (>140 mm Hg and/or 90 mm Hg) must be confirmed by home BP monitoring (>135 mm Hg and/or 85 mm Hg). Participants must be covered by or entitled to social security. A written informed consent must be obtained from the participants. Patient following any measures of legal presentation will be excluded, as will pregnant or breastfeeding women and women of childbearing age without a highly effective contraceptive measure (combined or progestogen-only hormonal contraception associated with inhibition of ovulation, intrauterine device or intrauterine hormone-releasing system). Participants will also be excluded if they present: clinical signs of hypovolaemia, symptomatic orthostatic hypotension, hyponatraemia (<130 mmol/L), dyskalemia (<3.5 mmol/L or >5.5 mmol/L), major adverse cardiovascular events during the last 3 months (myocardial infarction, heart failure hospitalisation, stroke), severe heart failure (New York Heart Association(NYHA) III_IV), current medical history of cancer requiring chemotherapy, solid organ transplantation or cirrhosis Child B–C. Participants with two or more diuretic agents (loop diuretic, thiazides and thiazide-like diuretics), mineralocorticoid receptor antagonists other than finerenone, autosomal dominant polycystic kidney disease treated with tolvaptan or contraindication to diuretics involved in the algorithm will be excluded from taking part in the trial. Included participants will not be allowed to be involved in pharmacological study during the entire follow-up period. Patients included in this study will not receive compensation. Transport costs related to the inclusion visit will be covered. Interventions In both groups, patients will be provided a package at the beginning of the study including an upper arm semiautomatic validated BP monitor and records with recommendations. The records will explain how to use the BP monitor and how to perform home BP monitoring (during three consecutive days, three measurements in the morning (before taking drugs) and in the evening (after dinner), taken in a quiet room after 5 min of rest, seated with back and arms supported). They will have a notebook to report the home BP monitoring and flyer about lifestyle interventions for patients with hypertension: salt restriction, alcohol consumption restriction, dietary changes (increase in consumption of vegetables, fresh fruits, fish, nuts and unsaturated fatty acids; decrease in consumption of red meat; consumption of low-fat dairy products), body weight control, regular aerobic exercise and smoking cessation. At each subsequent follow-up visit, if BP is not controlled (>135 mm Hg and/or 85 mm Hg) based on the results from home BP monitoring, the clinician will adjust the drug therapy according to the management decided by randomisation group. Intervention group In the intervention group, the clinician will adjust the drug therapy according to the antihypertensive algorithm based on diuretic agents. This algorithm includes prescription of thiazide or thiazide-like diuretics in patients with stage 4 CKD (eGFR <30 mL/min/1.73 m²). The diuretic agents that compose the algorithm are indapamide, hydrochlorothiazide, amiloride and furosemide (more details in online supplemental table 2). The algorithm is not disclosed here in order to prevent contamination between study groups. It will be fully described in the final report once the trial has been completed. A detailed presentation of the algorithm will be provided to centres randomised to the intervention group prior to trial initiation. Flyers detailing the strategy will be distributed to physicians in centres randomised to the intervention group. In addition, a dedicated study forum will be set up to support investigators and address any questions, particularly those related to the algorithm, with access rights tailored to each study group. The coordinating investigator will also be available to provide assistance if needed. The choice of medicinal products will be made by the investigators of each participating unit. This approach aims to allow each unit to administer treatments with which they are familiar. All products must be administered in accordance with their Summary of Product Characteristics and the recommendations of relevant learned societies. If a new step of the algorithm is not tolerated (clinically or biologically) without possibility of concomitant management (ie, supplementation or chelating agents for potassium), the dose of the drug will first be reduced if possible or discontinued. The algorithm will then proceed to the next appropriate step. No predefined criteria are specified for assessing clinical or biological tolerance; this assessment is left to the discretion of the clinician. The coordinating investigator can be contacted for assistance at any time. If it is not possible to follow the algorithm, the patient will remain in the study, even if all diuretics are discontinued. Control group The clinician will adapt the antihypertensive strategy according to his own standard of care which can be pharmacological or non-pharmacological therapies. Intervention concomitant care All concomitant cares are authorised. Minoxidil should be avoided in front of the risk of side effects: fluid retention and congestive heart failure. Of note, sodium glucose cotransporter-2 (SGLT2) inhibitors are tolerated and do not constitute contraindication. For other BP lowering therapeutics, CCB (eg, amlodipine or lercanidipine) are preferred, if not yet prescribed. Participants timeline The recruitment period will be 48 months. Patients will be followed for at least 24 months and up to 36 months, that is, the last included patients will be followed over a 2-year period, while the follow-up of patients included at the beginning of the study will be a 3 year period. The total duration of the study will be 72 months. Inclusion and baseline assessment Visit of pre-screening: patients will be selected in the nephrology centres during a routine consultation (table 1). Patients who provide a written and informed consent and meet the inclusion and exclusion criteria will be asked to: (1) perform home BP measurements with a validated device provided by the study team (three measures in the morning and in the evening during three consecutive days), (2) provide urinary test (24-hour urine collection: urinary sodium excretion, urinary creatinine excretion, proteinuria) and (3) provide a blood test (creatinine, sodium, potassium, urea, haemoglobin, haematocrit) and a pregnancy test only for those in the intervention group. For those in the control group, a pregnancy test will be carried out at the physician’s discretion depending on the treatment prescribed. Visit of pre-screening: patients will be selected in the nephrology centres during a routine consultation (table 1). Patients who provide a written and informed consent and meet the inclusion and exclusion criteria will be asked to: (1) perform home BP measurements with a validated device provided by the study team (three measures in the morning and in the evening during three consecutive days), (2) provide urinary test (24-hour urine collection: urinary sodium excretion, urinary creatinine excretion, proteinuria) and (3) provide a blood test (creatinine, sodium, potassium, urea, haemoglobin, haematocrit) and a pregnancy test only for those in the intervention group. For those in the control group, a pregnancy test will be carried out at the physician’s discretion depending on the treatment prescribed. An appointment for the inclusion visit (V0) will be scheduled. Visit of inclusion day 0—baseline: in both groups, patients will be included in the study after obtaining confirmation of uncontrolled BP, that is >135 mm Hg and/or 85 mm Hg (mean result from home BP measurements). Medical history, demographic data and concomitant medications will be collected. Participants will undergo a physical examination and assessment of quality of life with PROMIS-29 (Patient Reported Outcome Measurement Information System). Results from blood and urinary tests will be examined. Antihypertensive therapy intensification will be prescribed according to the randomisation group (standard of care management for antihypertensive therapy intensification, for the control group, or antihypertensive therapy intensification according to the study algorithm, for the intervention group). Visit of inclusion day 0—baseline: in both groups, patients will be included in the study after obtaining confirmation of uncontrolled BP, that is >135 mm Hg and/or 85 mm Hg (mean result from home BP measurements). Medical history, demographic data and concomitant medications will be collected. Participants will undergo a physical examination and assessment of quality of life with PROMIS-29 (Patient Reported Outcome Measurement Information System). Results from blood and urinary tests will be examined. Antihypertensive therapy intensification will be prescribed according to the randomisation group (standard of care management for antihypertensive therapy intensification, for the control group, or antihypertensive therapy intensification according to the study algorithm, for the intervention group). Table 1Schedule of enrolment, intervention and outcome measurementsStudy periodVisit of prescreeningVisit of inclusionFollow-upTimepointPre-day 0Day 0 baselineM3M6Every 6 months(2–3 years)Enrolment:Eligibility screeningXInformed consentXChecking of inclusion and non-inclusion criteriaXX (confirmed by home or ambulatory BP monitoring)RandomisationXInterventions: Intervention group: intensification strategy based on diuretics XXXX Control group: standard of care XXXXOutcomes:Primary outcome: kidney failure with replacement therapy, eGFR decline ≥40%, cardiovascular events or deathXXXSecondary outcomes Office and home BPXXXX Glomerular filtration rateXXXX Proteinuria (g/day) or proteinuria/creatininuria (g/g)XXXX Diuretic useXXXX Quality of life PROMIS-29XX(M12, M24±M36)Safety outcomesXXXX.BP, blood pressureeGFR, estimated glomerular filtration rate; PROMIS, Patient Reported Outcome Measurement Information System. Schedule of enrolment, intervention and outcome measurements .BP, blood pressureeGFR, estimated glomerular filtration rate; PROMIS, Patient Reported Outcome Measurement Information System. Follow-up assessments and visits Visit at 3 months, 6 months and every 6 months during the follow-up period: patient follow-up will be conducted according to usual clinical practice. Only home BP measurements and quality of life assessment are specific to the study. Visit at 3 months, 6 months and every 6 months during the follow-up period: patient follow-up will be conducted according to usual clinical practice. Only home BP measurements and quality of life assessment are specific to the study. In both groups, patients will undergo clinical examination: office BP measurement (including orthostatic hypotension assessment), weight, fluid overload assessment and signs of volume depletion. The patients will provide their home BP measurement results (three measures in the morning and in the evening during three consecutive days, the week before the visit). Results from urinary test (24-hour urine collection: urinary sodium excretion, urinary creatinine excretion, proteinuria) and blood test (creatinine, sodium, potassium, urea, haemoglobin and haematocrit) will be collected. For the intervention group, a blood pregnancy test will be performed. For the control group, a pregnancy test will be carried out at the physician’s discretion depending on the treatment prescribed. Quality of life will be assessed at the 12-month and 24±36-month visits. Any adverse event and detailed reporting of serious adverse events will be documented. If BP is not controlled (>135 mm Hg and/or 85 mm Hg), the clinician will adjust the drug therapy according to the randomisation group (standard of care management for antihypertensive therapy intensification, for control group, or antihypertensive algorithm based on diuretic agents, for intervention group). Monthly during the follow-up period: only for the intervention group, a monthly urine pregnancy test will be performed by each woman of childbearing during the clinical trial (except for the months when a blood pregnancy test is scheduled, that is, at each visit). Monthly during the follow-up period: only for the intervention group, a monthly urine pregnancy test will be performed by each woman of childbearing during the clinical trial (except for the months when a blood pregnancy test is scheduled, that is, at each visit). Last visit: the last visit will be performed between 24 months and 36 months. Last visit: the last visit will be performed between 24 months and 36 months. Randomisation Clusters will be randomly allocated to the two groups all at once. A 1:1 ratio will be used and the randomisation will be stratified, considering three strata. A specific stratum includes the two centres of Tours and Nancy, which have developed the algorithm, which is the key component of the intervention assessed. One of these two centres will be randomised in the intervention group whereas the other will be in the control group. The two other strata comprised academic nephrology centres and non-academic centres, respectively. Because all clusters will be randomised all at once, a simple method to generate the allocation sequence will be used, without considering blocks. The allocation sequence will be computer-generated by a statistician not involved in the recruitment or follow-up of patients. Assignments will be communicated to centres just after randomisation. Allocation concealment will be insured at the cluster level. However, because patients will be recruited after cluster had been randomised (figure 1), and because there is no possible blinding in the trial, there will no longer be allocation concealment at the patient level. This limit is well known and classical in many cluster randomised trials.37 Blinding The very nature of the intervention assessed prevents any form of blinding, neither for care providers, nor for patients. To prevent bias in measurement of the outcome, the primary outcome will be adjudicated by a centralised blinded committee. The adjudication committee will adjudicate all events using international criteria endorsed by major learnt societies.38–40 Experienced nephrologists, intensive care specialists, cardiologists and vascular neurologists from F-CRIN (French Clinical Research Infrastructure Network) Investigation Network Initiative-Cardiovascular and Renal Clinical Trialists (INI-CRCT), French Society of Hypertension and National College of Therapeutics Teachers will be invited to participate. Outcomes Primary outcome The primary outcome is a time-to-event composite outcome of: Kidney failure with replacement therapy: dialysis or renal transplantation.39 Kidney failure with replacement therapy: dialysis or renal transplantation.39 eGFR decline ≥40% (CKD-EPI formula). eGFR decline ≥40% (CKD-EPI formula). Cardiovascular events (myocardial infarction, heart failure hospitalisation, stroke).38 Cardiovascular events (myocardial infarction, heart failure hospitalisation, stroke).38 All-cause mortality. All-cause mortality. Cardiovascular events and kidney failure with replacement therapy will be adjudicated by a blinded committee. This endpoint is clinically relevant, based on a validated surrogate endpoint41 and three hard outcomes, rated as critically important by patients, caregivers and health professionals in the Standardised Outcomes in NephroloGy initiative.42 Secondary outcomes The secondary outcomes are: Each component of the composite primary outcome assessed separately. Each component of the composite primary outcome assessed separately. Change from baseline in systolic and diastolic BP at 3 months and 6 months, then every 6 months (with both home BP monitoring and office BP measurement). Change from baseline in systolic and diastolic BP at 3 months and 6 months, then every 6 months (with both home BP monitoring and office BP measurement). Proportion of patients with controlled BP at 2 years (BP <135 mm Hg and/or 85 mm Hg with home BP monitoring). Proportion of patients with controlled BP at 2 years (BP <135 mm Hg and/or 85 mm Hg with home BP monitoring). Change from baseline in glomerular filtration rate estimated by CKD-EPI formula at 3 months and 6 months, then every 6 months. Change from baseline in glomerular filtration rate estimated by CKD-EPI formula at 3 months and 6 months, then every 6 months. Change from baseline in proteinuria (g/day) or proteinuria/creatininuria (g/g) at 3 months and 6 months, then every 6 months. Change from baseline in proteinuria (g/day) or proteinuria/creatininuria (g/g) at 3 months and 6 months, then every 6 months. Proportion of patients who used at least one diuretic. Proportion of patients who used at least one diuretic. Change from baseline in quality of life assessed by PROMIS-29 survey each year.43 Change from baseline in quality of life assessed by PROMIS-29 survey each year.43 Safety (assessed over the minimal 2 years follow-up and up to 3 years):Volume depletion.44Dyskalemia: hypokalaemia (mild <3.5 mmol/L; moderate <3.0 mmol/L; severe <2.5 mmol/L), hyperkalaemia (mild >5.0 mmol/L; moderate >5.5 mmol/L; severe >6.0 mmol/L)Hyponatraemia <130 mmol/L.Acute kidney injury.45 46Gout.Orthostatic hypotension symptomatic/asymptomatic or complicated by falls. Safety (assessed over the minimal 2 years follow-up and up to 3 years): Volume depletion.44 Volume depletion.44 Dyskalemia: hypokalaemia (mild <3.5 mmol/L; moderate <3.0 mmol/L; severe <2.5 mmol/L), hyperkalaemia (mild >5.0 mmol/L; moderate >5.5 mmol/L; severe >6.0 mmol/L) Dyskalemia: hypokalaemia (mild <3.5 mmol/L; moderate <3.0 mmol/L; severe <2.5 mmol/L), hyperkalaemia (mild >5.0 mmol/L; moderate >5.5 mmol/L; severe >6.0 mmol/L) Hyponatraemia <130 mmol/L. Hyponatraemia <130 mmol/L. Acute kidney injury.45 46 Acute kidney injury.45 46 Gout. Gout. Orthostatic hypotension symptomatic/asymptomatic or complicated by falls. Orthostatic hypotension symptomatic/asymptomatic or complicated by falls. Sample size calculation Because the trial is cluster randomised, we must consider the intracluster correlation in the sample size calculation. Some approaches have been proposed, but they rely on an intracluster correlation value, which is a challenging issue for time-to-event outcomes. As a consequence, we specified the required sample size considering the outcome as a binary one. We assume that in the control group, the proportion of patients who will experience at least one component of the primary outcome within 2-year follow-up is 20%. This assumption is based on the CKD-REIN database. We hypothesise an 8% absolute decrease in the proportion of events in the intervention group (ie, 12% of events). In case the trial would have been an individually randomised trial, considering an 80% power with a two-sided type I error of 5%, we would have needed 326 patients per group considering the outcome as a binary one, and 333 per group, considering the outcome as a time-to-event one. At the design stage, we planned to have 36 centres (four additional centres were subsequently included). Considering an intracluster correlation of 0.005, we need to include 20 patients per cluster to obtain a power of 80.4%. This leads to a total number of 720 patients. Statistical analysis A statistical analysis plan will be finalised prior to database lock and the initiation of data analysis. All included patients will be analysed according to the cluster to which they were initially recruited, regardless of subsequent events. Similarly, all clusters will be analysed in the groups to which they were originally randomised. Baseline characteristics will be summarised by randomised groups in a table, using descriptive statistics. The number of participants with missing data for each variable of interest will be indicated. Analysis of the primary outcome The primary outcome is a composite of KFRT, eGFR decline ≥40%, cardiovascular events and all-cause mortality. Patients with missing primary outcome due to lost to follow-up will be censored at their last available assessment in the analysis. Survival will be described using Kaplan-Meier curves of time-to-first event. The main analysis will consist of a survival analysis of time to first event conducted with a Cox proportional hazard model adjusted on the stratification variable (type of centres, ie, academic vs non-academic centres). To account for within-cluster correlation of outcomes, the hazard ratio and 95% confidence interval (CI will be estimated with generalised estimating equation.47 If the proportional hazards assumption is not met, the primary analysis will rely on a difference in restricted mean survival time (RMST) at 24 months between the intervention and the control groups and its 95% CI estimated using pseudovalues regression.48 Given the number of randomised clusters (k=40), small-sample corrections will be used for estimating both RMST and HR. If the proportional hazards assumption holds, a difference in RMST will also be estimated to provide an absolute measure of the intervention effect as recommended in the Consolidated Standards of Reporting Trials (CONSORT) statement.49 Sensitivity analyses The Cox model described above will be rerun after adding the following patient’s baseline covariates to the model if there is a baseline imbalance (standardised mean difference >10%): age, sex, CKD stage 3b versus 4, diabetes, proteinuria, history of myocardial infarction and SGLT2. Other additional covariates could be added in function of potential baseline imbalances.50 This sensitivity analysis will allow to explore departure from the censoring completely at random assumption. We will perform a sensitivity analysis by excluding the two centres of Tours and Nancy which have been involved in the development of the algorithm. The intracluster correlation in each group will be estimated using Kendall’s τ. Analysis of secondary and safety outcomes Each component of the primary outcome will be assessed separately. All-cause mortality will be analysed using the same strategy as the primary outcome. Other time-to-event secondary outcomes (KFRT, eGFR decline ≥40% and cardiovascular events) will be analysed with competing risk methods accounting for within-cluster correlation of outcome, with death treated as a competing event.51 Continuous outcomes with repeated measurements, that is, BP, glomerular filtration rate, proteinuria, proteinuria/creatininuria and quality of life with PROMIS-29 will be analysed considering a mixed linear model with two random effects: one associated to patients (thus taking into account intra-patient correlation) and the other to clusters (thus taking into account the intracluster correlation). Binary secondary outcomes, that is, percentage of patients with controlled BP at 2 years and proportion of patients who used at least one diuretic, will be analysed considering a marginal approach by fitting a linear regression with Gaussian regression with robust standard errors within a generalised estimating equation framework.52 Safety outcomes will be reported and expressed as proportions in each group and compared using the same strategy as the binary secondary outcomes. Analysis will be performed in SAS V.9.4 (SAS institute, Cary, North Carolina, USA) and R V.3.3.1 software (or latest versions). Data collection, management and monitoring All protocol-required information will be recorded in the electronic case report form. To ensure participants’ confidentiality, each form will be identified by a unique patient identification number and initials. Investigators will receive standardised training to ensure consistent and reliable data entry and will be responsible for the accuracy and completeness of the collected data. Data will be entered at the investigational sites via a secure web-based platform, under the responsibility of the investigator. A data manager will perform regular data monitoring and issue queries to resolve missing, inconsistent or implausible data. The database will be locked after all queries have been resolved and a blinded data review has been completed, prior to statistical analysis. Consistent with the pragmatic design of the study, no formal verification of the accuracy home BP monitoring measurements will be performed. BP measurements will be self-reported by participants as they would in routine clinical practice. Given the nature of the intervention (antihypertensive medicinal products used in accordance with their marketing authorisation, the recommendations of learnt societies), no data safety monitoring committee is planned for this clinical trial. Ethics and dissemination Confidentiality of data In compliance with applicable regulations (EU Regulation 2016/679—General Data Protection Regulation (GDPR), and its transposition into French law via Law No. 2018–493 of 20 June 2018, from the Code de Santé Publique), individuals with direct access to source data will take all necessary measures to safeguard the confidentiality of information related to the study, its participants and the results obtained, particularly with regard to participants’ identity. These individuals, including investigators, are bound by professional confidentiality. During the study and after its completion, data transmitted to the sponsor by investigators or qualified personnel will be coded to protect participants’ identity. The sponsor will ensure that each participant has provided written consent for access to their personal data, strictly limited to what is necessary for study quality control. Ethical consideration and approval The protocol, informed consent form and patient information notice (online supplemental material 3) were reviewed and approved by a French protection of individuals committee (CPP) prior to study initiation (CPP Ile-de-France VII, 2022–5 01 494-39-00, 23 January 2023). An agreement from the French National Agency for Security of Medicines and Healthcare Products (ANSM) has also been obtained. The study will be conducted in compliance with the French law in force (Code de Santé Publique), the recommendations of French and international Good Clinical Practices, the Helsinki Declaration (Ethical Principles for Medical Research involving Human Subjects) and the European regulations related to clinical research (Regulation (EU) No 536/2014 of the European Parliament and of the Council of 16 April 2014 on clinical trials on medicinal products for human use). The data collected during the trial will be processed in accordance with the requirements of the National Commission on Informatics and Liberty(CNIL) and GDPR. General Data Protection Regulation GDPR. An audit may be conducted at any time by individuals appointed by the sponsor who are independent of the study team. The audit may cover all phases of the study, including protocol development, data collection and storage, result analysis and publication. Protocol amendments Any substantial amendment affecting participant safety, study validity, investigational product quality or interpretation of scientific data must be submitted in writing to the sponsor before implementation. Approval from the ethics committee and authorisation from the ANSM are required. If the amendment alters participant management or changes the risks, benefits or constraints, a revised patient information form and informed consent form must be provided. All investigators involved in the study must be informed of any protocol amendments. Dissemination Results from this study will be submitted for publication in peer-reviewed scientific journals. We will follow the Recommendations for the Conduct, Reporting, Editing and Publication of Scholarly Work in Medical Journals (updated in January 2026) from the International Committee of Medical Journal Editors and comply with the CONSORT statement (http://www.consort-statement.org) and the extension to cluster randomised trials in the publication of the main results. Any person with a key contribution to the study will be considered as authors. Results will also be submitted for presentation in international and national scientific conferences. In accordance with the regulatory provisions in force, participants will be informed, at their request, of the overall results of the study. Also, the summary of results will be accessible via the European platform (https://euclinicaltrials.eu/search-for-clinical-trials). Data sharing Following publication of primary results, deidentified data will be made available from the principal investigator on any reasonable request and based on a methodologically sound proposal. The conditions for the transfer of all or part of the database will be decided by the study sponsor (University Hospital of Tours) and will be the subject of a written contract. Data dictionaries, study protocol, statistical analysis plan and informed consent form could be shared under the same conditions. Patient and public involvement A patient expert from INI-CRCT will be engaged in the steering committee. Study status The first patient was recruited in the THINK trial on 28 March 2023. Recruitment is expected to continue until 28 March 2027 with a follow-up completed on 28 March 2029. The current protocol is V.3, 13 March 2025. Discussion Hypertension is the first modifiable risk factor for cardiovascular events and kidney disease progression in CKD. However, despite guideline recommendations, optimal antihypertensive sequencing in stages 3b–4 CKD remains uncertain, as few randomised trials have evaluated treatment strategies specifically in this population. Real-world data consistently show high rates of uncontrolled and resistant hypertension, together with a major contribution of extracellular volume overload, while the use of diuretics remains heterogeneous across nephrology practice. The THINK trial addresses this gap by evaluating a structured, diuretic-based BP-lowering algorithm compared with standard of care in patients with moderate to advanced CKD. The underlying rationale builds on strong physiological mechanisms and observational evidence showing that volume overload is a key driver of uncontrolled BP, cardiovascular risk and progression to kidney failure. By testing an algorithm combining diuretics with ACEi or ARB therapy, the trial aims to determine whether this strategy can improve major clinical outcomes including KFRT, cardiovascular events and mortality. The pragmatic cluster-randomised design is a major strength, enhancing generalisability and minimising treatment contamination. The participation of both academic and non-academic centres increases representativeness, and the use of home BP monitoring ensures reliable measurements for therapeutic adjustment. Furthermore, clinical events will be adjudicated by an independent blinded committee, strengthening the internal validity of outcome assessment. The study has some limitations. Blinding is not feasible due to the nature of the intervention, which may introduce bias in the identification and recruitment of the patients, and measurement bias, although the use of objective and adjudicated outcomes mitigates this risk. Cluster randomisation may lead to baseline imbalances, inherent to this design, but will be addressed analytically. Diuretic-based intensification may increase the risk of electrolyte disorders or acute kidney injury but these safety outcomes will be closely monitored throughout follow-up. In summary, the THINK trial is designed to provide robust evidence regarding a diuretic-based strategy for patients with stages 3b–4 CKD. If effective, this approach could inform future clinical guidelines and contribute to improving cardiovascular and renal outcomes in patients with CKD. Supplementary material