Pharmacological blood-pressure lowering for the prevention of cardiovascular

# Pharmacological blood-pressure lowering for the prevention of cardiovascular disease and death across the full spectrum of chronic kidney disease severity: an individual-participant data meta-analysis *The Lancet 2026* Articles Pharmacological blood-pressure lowering for the prevention of cardiovascular disease and death across the full spectrum of chronic kidney disease severity: an individual-participant data meta-analysis Guyu Zeng, Zeinab Bidel, Qianqian Yang, Dexter Canoy, Mark Woodward, Julia Lewis, Sverre E Kjeldsen, William C Cushman, Jinqing Yuan, Koon Teo, Barry R Davis, John Chalmers, Carl J Pepine, Kazem Rahimi, Milad Nazarzadeh, on behalf of the Blood Pressure Lowering Treatment Trialists’ Collaboration* Summary Lancet 2026; 407: 1626–38 Background Individuals with chronic kidney disease (CKD), particularly those at more advanced stages, have been See Comment page 1578 systematically under-represented in randomised controlled trials (RCTs) of blood-pressure-lowering treatment due to *Members listed in the appendix safety concerns, leading to a persistent paucity of evidence for cardiovascular risk management in this high-risk (p 4) group. We investigated the eect of blood-pressure-lowering treatment on the risk of major cardiovascular disease Deep Medicine Group, Nuffield and death across the full spectrum of CKD stages and by key clinical subgroups. Department of Women’s and Reproductive Health, Medical Methods We conducted a one-stage meta-analysis of individual-participant data from RCTs in which participants were Sciences Division, University of randomly assigned to a blood-pressure-lowering therapy versus a comparator. We used RCTs collated in the Blood Oxford, Oxford, UK (G Zeng MD, Z Bidel MSc, Q Yang MSc, Pressure Lowering Treatment Trialists’ Collaboration dataset, published at any time in any language, which were Prof M Woodward PhD, eligible for inclusion if they had at least 1000 person-years of follow-up per arm, baseline blood-pressure and creatinine Prof K Rahimi FRCP, measurements, and time-to-event outcomes; those with unclear randomisation procedures or restricted to heart M Nazarzadeh DPhil); failure or acute care settings were excluded. Participants with a documented history of heart failure or extreme Department of Cardiology, Fuwai Hospital, National Center creatinine values were excluded. No age criteria were applied. The primary outcome was major cardiovascular events, for Cardiovascular Disease, defined as a composite of fatal or non-fatal stroke, ischaemic heart disease, or hospitalisation for, or death from, heart Peking Union Medical College, failure. Relative treatment eects were estimated with a stratified Cox proportional hazards model. Heterogeneity of Chinese Academy of Medical Sciences, Beijing, China (G Zeng, treatment eects was evaluated across prespecified subgroups defined by CKD status, CKD stage (1–5), diabetes, Prof J Yuan PhD); Population proteinuria, and baseline blood pressure. A stratified network meta-analysis was performed to examine whether Health Sciences Institute, treatment eects diered by defined subgroups within each of five principal antihypertensive drug classes. The Newcastle University, systematic review was registered in PROSPERO (CRD42018099283). Newcastle, UK (D Canoy PhD); The George Institute for Global Health, School of Public Health, Findings From 52 RCTs (363 684 participants), a total of 285 124 participants from 46 randomised trials met the Imperial College London, eligibility criteria; 116 145 (40·7%) were women, 168 979 (59∙3%) were men, 59 185 (20·7%) had CKD at baseline, and London, UK (Prof M Woodward); 86 067 (30·2%) had type 2 diabetes. During a median follow-up of 4·4 years (IQR 3·2–5·1), a 5 mm Hg reduction in The George Institute for Global Health, University of New systolic blood pressure reduced the risk of major cardiovascular disease in individuals with CKD (hazard ratio South Wales, Sydney, NSW, [HR] 0·91 [95% CI 0·87–0·94]) and without CKD (0·90 [0·88–0·93]; p >0·99). Furthermore, these observed interaction Australia (Prof M Woodward, relative risk reductions were consistent across all CKD stages, including severe stages 4–5 (p >0·99). Similar Prof J Chalmers PhD); Division of interaction treatment eects were observed by proteinuria status and across blood-pressure categories, down to <120/70 mm Hg. Nephrology and Hypertension, Department of Medicine, However, the relative treatment eect in individuals with CKD was notably attenuated among those with coexisting Vanderbilt University Medical diabetes (HR 0·96 [95% CI 0·90–1·02]) compared with those without (0·88 [0·84–0·93]; p =0·044). The stratified interaction Center, Nashville, TN, USA analysis within each drug class showed that the class-specific eects of antihypertensive agents versus placebo on (Prof J Lewis MD); Department cardiovascular disease risk remained unchanged across the investigated subgroups. of Cardiology, Department of Nephrology, and Institute of Clinical Medicine, University of Interpretation In the context of cardiovascular risk reduction, the relative benefit of blood-pressure lowering in Oslo, Ullevaal Hospital, Oslo, patients with CKD is similar to that in individuals without CKD, with consistent ecacy across all CKD stages, blood- Norway (Prof S E Kjeldsen MD); pressure thresholds, and proteinuria status. However, notably, this relative benefit is attenuated in patients with CKD Department of Preventive Medicine, The University of and concomitant diabetes, underscoring the requirement for adapted therapeutic strategies in this high-risk subgroup. Tennessee Health Science Moreover, the class-specific eects of principal antihypertensives in CKD mirror those observed in the broader Center, Memphis, TN, USA population, independent of CKD stage or proteinuria status. (Prof W C Cushman MD); Population Health Research Institute, Hamilton Health Funding British Heart Foundation. Sciences, McMaster University, Hamilton, ON, Canada Copyright © 2026 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 (Prof K Teo PhD); The University license. of Texas School of Public Health, Houston, TX, USA 1626 Articles (Prof B R Davis PhD); College of Research in context Medicine, University of Florida, Gainesville, FL, USA Evidence before this study or higher disease progression, allowing for a comprehensive (Prof C J Pepine MD) We searched PubMed and the Cochrane Library from database assessment across the entire spectrum of CKD. We found that a Correspondence to: inception to Jan 31, 2025, using MeSH terms and keywords for 5 mm Hg reduction in systolic blood pressure is associated with Dr Milad Nazarzadeh, Deep “hypertension”, “blood pressure”, “chronic kidney disease”, a relative risk reduction in major cardiovascular disease in both Medicine Group, Nuffield and “antihypertensive agents”, including variant terms and CKD and non-CKD participants, with consistent treatment Department of Women’s and Reproductive Health, Medical relevant drug classes, without language restrictions. Existing effects across all CKD stages and baseline blood-pressure values, Sciences Division, University of evidence on the cardiovascular benefits of blood-pressure and by proteinuria status. We found that the relative treatment Oxford, Oxford OX2 0EW, UK lowering in chronic kidney disease (CKD) is scarce, effect is attenuated in participants with coexisting diabetes, milad.nazarzadeh@wrh.ox.ac. inconsistent, and often non-generalisable, deriving largely highlighting a crucial subgroup of patients with CKD for uk from individual trials that enrolled participants with mildly targeted treatment strategies. Furthermore, the class-specific See Online for appendix reduced kidney function or were underpowered to examine effect of antihypertensives versus placebo was similar in those treatment effects across the full spectrum of disease severity. with CKD compared with those without CKD, with similar No previous meta-analysis has comprehensively assessed effects across all stages and by proteinuria status. treatment effects across all CKD stages—particularly stages Implications of all the available evidence 4–5—or examined effect modification by diabetes and When the main therapeutic goal is cardiovascular risk proteinuria within a unified individual-participant data management, clinicians can recommend blood-pressure- framework. Studies of class-specific antihypertensive effects lowering treatment to individuals at all stages of CKD have predominantly focused on kidney outcomes, leaving progression, regardless of blood-pressure values, provided the uncertainty about whether the cardiovascular efficacy of balance of benefits and harms is favourable and patient individual drug classes observed in the general population is preferences are considered. This recommendation can be done maintained in people with CKD, and whether any class confers with the expectation that the class-specific effects of different differential cardiovascular benefit across disease stages or by antihypertensive drugs mirror those observed in the broader proteinuria status. population. In individuals with CKD and coexisting diabetes, Added value of this study blood-pressure-lowering treatment is essential due to their This one-stage, individual-participant data meta-analysis, notably elevated absolute risk and the well documented pooling data from 46 large-scale trials involving beneficial effects of treatment in patients with diabetes. 285 124 participants (20·7% with CKD at baseline), represents However, the attenuated relative risk reduction associated with the largest randomised dataset to date in this population. diabetes highlights the need for adapted strategies to enhance Importantly, 14 148 (23·9%) of the CKD subgroup had stage 3b cardiovascular risk management in this population at high risk. Introduction eects; and (4) and whether specific classes of The cardioprotective benefits of blood-pressure-lowering antihypertensive drugs dier in their ability to reduce therapy are well documented across diverse populations cardiovascular risk in CKD, depending on disease stage with varied clinical backgrounds.1–3 However, the ecacy or the presence or absence of proteinuria.7 and optimal application of blood-pressure-lowering Individual blood-pressure-lowering RCTs have not therapy in people with chronic kidney disease (CKD) succeeded in bridging this gap and, in some cases, have remain insuciently investigated. This evidence gap further complicated the evidence base, thereby making arises primarily from the under-representation of patients clinical interpretation more challenging. For example, with CKD in randomised controlled trials (RCTs) due to although earlier trials did not report a cardiovascular concerns about kidney-related harm that have led to their benefit of blood-pressure lowering in individuals with frequent exclusion from blood-pressure-lowering trials.4 CKD,8 the Systolic Blood Pressure Intervention Trial As CKD severity increases, evidence becomes increasingly (SPRINT)9,10 showed a transparent and similar reduction scarce, especially in later stages of the disease, leaving in risk with intensive blood-pressure lowering in both clinicians dependent on data from non-CKD or lower-risk CKD and non-CKD participants. However, a post-hoc populations.5 Despite widespread recognition of this analysis of SPRINT10 suggested a decreasing trend in the evidence gap, progress over the past two decades has relative reduction of risk with progression to more been slow6 and several crucial questions remain: (1) advanced CKD stages. Moreover, trials published in 2025 whether blood-pressure-lowering ecacy varies by CKD explicitly conducted in patients with advanced CKD have status, stage, or baseline blood pressure; (2) whether not shown an apparent reduction in cardiovascular risk blood-pressure-lowering treatment influences the risk of with antihypertensive treatment.11,12 cardiovascular disease or death in patients with advanced Similarly, meta-analyses of RCTs have produced CKD, potentially conferring either benefit or harm; (3) inconsistent findings. An aggregate data meta-analysis of whether diabetes and proteinuria modify treatment 18 RCTs showed that a 10 mm Hg reduction in blood Articles pressure decreased major cardiovascular events in both requiring that all studies are prespecified and approved the CKD and non-CKD groups; however, the benefit was before data are released for analysis. The study obtained smaller in the CKD group.13 An individual-participant ethics approval from the Oxford Tropical Research Ethics data meta-analysis of 23 RCTs with 152 290 participants Committee (reference 545–14), and each contributing indicated that blood-pressure lowering reduced trial secured informed consent from its participants.15 cardiovascular risk in both groups, with no advantage For our meta-analysis, we included BPLTTC trials that related to specific drug classes.14 However, most provided data on baseline blood pressure, baseline participants with CKD (76%) had estimated glomerular creatinine values, major cardiovascular events, cause- filtration rate (eGFR) values of 45–60 mL/min per specific cardiovascular and all-cause death, and the 1·73 m² and only 0·4% had eGFRs lower than 30 mL/min corresponding dates of occurrence. We excluded per 1·73 m². Limitations of this meta-analysis, such as participants with a history of heart failure and those with varying treatment definitions, the absence of a direct extreme creatinine values (ie, <0·2 mg/dL or >5·0 mg/dL). interaction assessment, the absence of standardisation Treatment and comparator groups were defined for blood-pressure reduction, and the small number of according to the original trial design. In placebo- trials included, hampered definitive conclusions.14 controlled trials, the active treatment arm was considered We aimed to address these gaps by pooling individual- the intervention and the placebo arm was considered the participant data from large-scale RCTs representing the comparator. In head-to-head trials comparing drug largest known randomised dataset for this population that classes, the arm with the greater reduction in systolic included, importantly, participants with eGFR below blood pressure was designated the intervention, and the 45 mL/min per 1·73 m² and 30 mL/min per 1·73 m² or arm with the lesser reduction was designated the lower. comparator. In trials with dierent treatment intensities, the intensive arm was the intervention and the standard Methods arm was the comparator. Study design and procedures We estimated eGFR using the CKD Epidemiology In this individual-participant data meta-analysis, we used Collaboration 2021 race-free equations (appendix p 5).18 RCTs from the third cycle of the Blood Pressure Lowering For subgroup analyses by baseline CKD status, Treatment Trialists’ Collaboration (BPLTTC).15 The participants with an eGFR less than 60 mL/min per BPLTTC, with the first cycle established in 1995, is an 1·73 m² were categorised as having CKD at baseline.7,19 international collaboration of investigators from major The presence of proteinuria was defined as a protein-to- RCTs of pharmacological blood-pressure-lowering creatinine ratio of 0·22 or higher, urinary albumin therapies, dedicated to assessing their eects across excretion 200 μg/min or higher (>300 mg/day), urinary diverse populations and clinical subgroups.16 The general albumin concentration of 200 mg/L or higher, urinary eligibility criteria for inclusion in the BPLTTC dataset albumin-to-creatinine ratio of 30 mg/g or higher, or a were RCTs, published at any time and in any language, urinary protein dipstick result of 1 or higher.20,21 To assess that compared an antihypertensive drug versus placebo the eects of blood-pressure-lowering treatment across or another antihypertensive, or investigated dierent the spectrum of CKD severity, five stages were defined blood-pressure-lowering intensities, with a minimum of according to baseline eGFR: stage 1 (≥90 mL/min per 1000 patient-years of follow-up in each randomised arm. 1·73 m²); stage 2 (60–89 mL/min per 1·73 m²); stage 3a There were no age criteria for participant inclusion. (45–59 mL/min per 1·73 m²); stage 3b (30–44 mL/min Trials without a clearly defined randomisation process, per 1·73 m²); and stages 4–5 (<30 mL/min per 1·73 m²).22 those evaluating non-pharmacological interventions, and Baseline systolic and diastolic blood pressure those conducted exclusively in patients with heart failure measurements were categorised by 10 mm Hg intervals, or in short-term acute settings (such as acute myocardial resulting in seven categories for systolic blood pressure infarction) were excluded.1,15 Further methodological (ranging from <120 mm Hg to ≥170 mm Hg) and details, including the central systematic review, quality six for diastolic blood pressure (ranging from <70 mm Hg and risk-of-bias assessments, characteristics of the to ≥110 mm Hg). Baseline diabetes status was established included trials, and estimated achieved blood-pressure based on the diagnostic information provided by each trial.3 reductions, have been reported previously1,15,17 and are Artificial intelligence (AI) was not used in the design, available in the appendix (pp 16–17, 52). conduct, analysis, or interpretation of this study. However, The BPLTTC operates in accordance with the University AI-assisted tools were used for language editing, of Oxford’s policies on research integrity, codes of proofreading, and R code debugging. practice, and the management of research data and records. The systematic review protocol underpinning Data analysis the third cycle of BPLTTC, specifying eligibility criteria, The primary outcome was the first occurrence of a major search strategy, and analytical methods, was prospectively cardiovascular disease, defined as a composite of fatal or registered in PROSPERO (CRD42018099283). The non-fatal stroke or other cerebrovascular disease, fatal or Steering Committee oversees all scientific activities, non-fatal ischaemic heart disease, or heart failure leading 1628 Articles to death or hospitalisation. Secondary outcomes included overall eect was considered the most valid estimate of the individual components of the primary outcome as the treatment eect. A significant interaction was well as cardiovascular and all-cause death. We defined interpreted in the context of eect magnitude and and ascertained outcomes based on the diagnostic direction, previous literature, biological plausibility, and definitions and endpoint adjudication criteria applied in clinical relevance. each contributing trial, using the adjudicated event data In addition to the main analysis, we conducted an supplied in the individual-participant datasets. individual-participant data network meta-analysis to All core variables, including comparison arms, blood- estimate stratified, class-specific treatment eects for the pressure measurements after treatment, trial endpoints, five principal antihypertensive drug classes: angiotensin- general baseline characteristics including sex, and converting enzyme inhibitors, angiotensin receptor diabetes status, had already been harmonised in previous blockers, β blockers, calcium channel blockers, and BPLTTC studies and were used in this analysis.1–3,17,23,24 In thiazide diuretics.3,23 This analysis was designed to assess parallel, a dedicated harmonisation process was done for whether eects within drug classes varied by CKD status, variables specific to this study: baseline creatinine, eGFR, CKD stage, or presence of proteinuria, which is a and presence or absence of proteinuria. We applied a research question that is not feasible to investigate with a fixed-eects, one-stage, individual-participant data meta- conventional network meta-analysis. The comparison or analysis framework, pooling participant-level data from all ranking of drug classes was not the objective of this study eligible trials and analysing them as a single, large-scale because this question has been investigated previously in dataset, with each trial as a cluster. several aggregate-data meta-analyses.13,30,31 Logistic We fitted Cox proportional hazards models, stratified regression models were applied to individual-level data by trial, to allow trial-specific baseline hazards and to from each trial to estimate the odds ratio as the relative control for between-trial dierences in baseline risk.25 treatment eect for each available comparison, stratified Given the design of the included trials, the main source by CKD status, CKD stage, and proteinuria status. of heterogeneity was the variation in blood-pressure Finally, the estimated eects were pooled with a fixed- reduction after treatment, driven primarily by eect Bayesian network meta-analysis based on Markov dierences in comparison arms and treatment types.17 Chain Monte Carlo methods (four chains, 10 000 burn-in To account for this heterogeneity, the models were iterations, and 100 000 sampling iterations), with the standardised for the blood-pressure reduction after placebo arm serving as the network reference.32 To assess treatment at the trial level, and hazard ratios (HRs) were whether drug-class-specific eects varied by subgroups, rescaled to express the relative treatment eect per Wald-type Z tests were used to compare subgroup- 5 mm Hg reduction in systolic blood pressure and specific log-odds ratios.33 For each drug class, the linear 3 mm Hg reduction in diastolic blood pressure (ie, trend across CKD stages was assessed with meta- values representing the mean reductions reached in all regression with CKD stage as an ordinal covariate, BPLTTC trials, excluding head-to-head RCTs; appendix deriving the p value for trend from the test of p 6).1,2,17 This parametrisation implicitly scales each moderators.34 trial’s contribution by the blood-pressure reduction after All analyses in this study were conducted according to treatment such that trials with larger reductions the intention-to-treat principle. We performed statistical contribute more information to the standardised eect analyses using R (version 4.2.0). Details of packages used estimate, whereas trials with smaller reductions are for analysis are reported in the appendix (p 15). retained but contribute correspondingly less. This approach prevents the arbitrary exclusion of trials with Role of the funding source small blood-pressure reductions after treatment, The funder of the study had no role in study design, data maximises statistical power for subgroup analyses, and collection, data analysis, data interpretation, or writing of yields eect estimates with clear clinical interpretation the report. (appendix pp 7–12).26 The cumulative probability of major cardiovascular disease was estimated in each Results treatment arm using the Kaplan–Meier method and The current third cycle of BPLTTC included individual- plotted separately for subgroups with and without CKD level data from 52 RCTs, encompassing and by CKD stage. 363 684 participants.15 From these RCTs, we excluded In subgroup analyses, the likelihood-ratio test was used one trial due to the absence of time-to-event data35 and to assess interactions between treatment and subgroups, five trials for not having baseline creatinine measure- with p values for interaction corrected for multiple ments.36–40 Consequently, 46 trials comprising comparisons by use of Hommel’s method (appendix 285 124 participants (59 185 with CKD and 225 939 without) pp 11–12).27 Results from subgroup analyses were met the eligibility criteria and were included in the reported and interpreted in accordance with established analysis (appendix pp 17–24). A detailed flow chart principles for clinical trial interpretation.28,29 When the showing the study selection is available in the appendix statistical test for interaction was non-significant, the (p 35). Baseline diabetes status was available for all Articles Individuals with CKD at baseline (n=59 185) Individuals without CKD at baseline (n=225 939) Intervention Comparator Total Intervention Comparator Total Sex Female 14 340/27 804 (51·6%) 16 201/31 381 (51·6%) 30 541/59 185 (51·6%) 40 474/106 149 (38·1%) 45 130/119 790 (37·7%) 85 604/225 939 (37·9%) Male 13 464/27 804 (48·4%) 15 180/31 381 (48·4%) 28 644/59 185 (48·4%) 65 675/106 149 (61·9%) 74 660/119 790 (62·3%) 140 335/225 939 (62·1%) Age, years 69·5 (9·4) 70·0 (9·2) 69·8 (9·3) 63·6 (9·6) 64·1 (9·3) 63·9 (9·5) Systolic blood pressure, 156 (23) 155 (23) 156 (23) 153 (21) 152 (21) 152 (21) mm Hg Diastolic blood pressure, 86 (13) 86 (13) 86 (13) 88 (13) 88 (12) 88 (13) mm Hg BMI, kg/m² 28·1 (5·2) 28·1 (5·6) 28·1 (5·4) 27·8 (5·0) 28·0 (9·9) 27·9 (8·0) Smoking status Never 10 082/20 001 (50·4%) 11 366/23 019 (49·4%) 21 448/43 020 (49·9%) 33 991/74 382 (45·7%) 37 972/86 375 (44·0%) 71 963/160 757 (44·8%) Past 7108/20 001 (35·5%) 8416/23 019 (36·6%) 15 524/43 020 (36·1%) 25 526/74 382 (34·3%) 31 463/86 375 (36·4%) 56 989/160 757 (35·5%) Current 2811/20 001 (14·1%) 3237/23 019 (14·1%) 6048/43 020 (14·1%) 14 865/74 382 (20·0%) 16 940/86 375 (19·6%) 31 805/160 757 (19·8%) Ethnicity White, Caucasian, or 13 837/21 202 (65·3%) 15 472/24 590 (62·9%) 29 309/45 792 (64·0%) 52 596/80 312 (65·5%) 60 554/92 425 (65·5%) 113 150/172 737 (65·5%) European Black 2893/21 202 (13·6%) 3856/24 590 (15·7%) 6749/45 792 (14·7%) 4852/80 312 (6·0%) 7241/92 425 (7·8%) 12 093/17 737 (68·2%) Hispanic 965/21 202 (4·6%) 1386/24 590 (5·6%) 2351/45 792 (5·1%) 3751/80 312 (4·7%) 5644/92 425 (6·1%) 9395/17 737 (53·0%) Asian 2484/21 202 (11·7%) 2792/24 590 (11·4%) 5276/45 792 (11·5%) 16 670/80 312 (20·8%) 16 452/92 425 (17·8%) 33 122/17 737 (186·7%) Other 1023/21 202 (4·8%) 1084/24 590 (4·4%) 2107/45 792 (4·6%) 2443/80 312 (3·0%) 2534/92 425 (2·7%) 4977/17 737 (28·1%) Comorbidity Peripheral vascular disease 1510/11 314 (13·3%) 1451/11 038 (13·1%) 2961/22 352 (13·2%) 3415/41 368 (8·3%) 3540/39 850 (8·9%) 6955/81 218 (8·6%) Atrial fibrillation 1300/14 345 (9·1%) 1404/16 519 (8·5%) 2704/30 864 (8·8%) 3124/55 179 (5·7%) 3373/63 051 (5·3%) 6497/118 230 (5·5%) Cerebrovascular disease 4466/22 080 (20·2%) 4879/23 813 (20·5%) 9345/45 893 (20·4%) 15 702/85 469 (18·4%) 17 132/91 583 (18·7%) 32 834/177 052 (18·5%) Ischaemic heart disease 8191/25 157 (32·6%) 9333/26 788 (34·8%) 17 524/51 945 (33·7%) 27 243/94 999 (28·7%) 31 934/101 237 (31·5%) 59 177/196 236 (30·2%) Type 2 diabetes 8474/27 801 (30·5%) 9582/31 377 (30·5%) 18 056/59 178 (30·5%) 31 824/106 109 (30·0%) 36 187/119 741 (30·2%) 68 011/225 850 (30·1%) Previous use of non-trial medications Diuretics 5210/15 556 (33·5%) 5967/16 673 (35·8%) 11 177/32 229 (34·7%) 9492/56 753 (16·7%) 11 054/61 340 (18·0%) 20 546/118 093 (17·4%) α blockers 833/12 126 (6·9%) 898/13 323 (6·7%) 1731/25 449 (6·8%) 1379/40 945 (3·4%) 1633/46 040 (3·5%) 3012/86 985 (3·5%) β blockers 5734/16 365 (35·0%) 6637/17 461 (38·0%) 12 371/33 826 (36·6%) 18 464/60 339 (30·6%) 21 748/64 936 (33·5%) 40 212/125 275 (32·1%) Angiotensin-converting 5780/14 954 (38·7%) 6699/16 105 (41·6%) 12 479/31 059 (40·2%) 14 827/53 623 (27·7%) 18 617/58 153 (32·0%) 33 444/111 776 (29·9%) enzyme inhibitors Angiotensin-receptor 748/9193 (8·1%) 746/8726 (8·5%) 1494/17 919 (8·3%) 3198/39 354 (8·1%) 3195/37 053 (8·6%) 6393/76 407 (8·4%) blockers Calcium-channel blockers 5977/16 377 (36·5%) 6360/17 475 (36·4%) 12 337/33 852 (36·4%) 18 634/60 342 (30·9%) 19 961/64 936 (30·7%) 38 595/125 278 (30·8%) Antiplatelets 4825/12 353 (39·1%) 6088/13 556 (44·9%) 10 913/25 909 (42·1%) 16 138/41 257 (39·1%) 21 601/46 943 (46·0%) 37 739/88 200 (42·8%) Anticoagulants 765/8027 (9·5%) 878/9188 (9·6%) 1643/17 215 (9·5%) 1804/26 857 (6·7%) 2221/32 457 (6·8%) 4025/59 314 (6·8%) Lipid-lowering treatments 4159/12 528 (33·2%) 5065/12 991 (39·0%) 9224/25 519 (36·1%) 16 540/50 834 (32·5%) 20 677/53 476 (38·7%) 37 217/104 310 (35·7%) eGFR, mL/min per 1·73m² 49·8 (8·6) 49·9 (8·6) 49·9 (8·6) 81·8 (13·5) 81·6 (13·3) 81·7 (13·4) Proteinuria 2975/14 036 (21·2%) 3023/15 152 (20·0%) 5998/29 188 (20·5%) 5434/46 821 (11·6%) 6146/52 076 (11·8%) 11 580/98 897 (11·7%) Follow-up, years 4·3 (3·0–5·0) 4·4 (3·0–5·0) 4·4 (3·0–5·0) 4·4 (3·2–5·1) 4·4 (3·3–5·1) 4·4 (3·3–5·1) Data are n/N (%), mean (SD), or median (IQR). Sex refers to biological sex as recorded by trial investigators at enrolment and harmonised across trials. Data on gender identity and psychosocial or cultural gender constructs were not available in the Blood Pressure Lowering Treatment Trialists’ Collaboration database and therefore could not be analysed. CKD=chronic kidney disease. eGFR=estimated glomerular filtration rate. Table: Baseline characteristics of participants by CKD status and study arm included trials and proteinuria measurements were at baseline for all included participants are presented in obtained from 24 trials (appendix pp 17–24). Compared the appendix (p 36). with those without CKD, patients with CKD were more During a median follow-up of 4·4 years (IQR 3·2–5·1), likely to be female and older, had higher baseline systolic the primary composite outcome occurred in 36 473 (12∙8%) blood pressure, were less likely to be smokers, and had a of 284 134 participants, with individual event counts of greater prevalence of cardiovascular comorbidities (table). ischaemic heart disease (17 817 [6∙3%] of 284 333), stroke The distribution of eGFR and corresponding CKD stages (12 795 [4∙5%] of 284 350), heart failure (6875 [2∙8%] of 1630 Articles 246 202), cardiovascular death (10 044 [3∙6%] of 282 222), (n=12 988; appendix pp 32–35). The analysis comparing and all-cause death (25 197 [8∙9%] of 284 365). In individual drug classes with placebo showed that class- participants with CKD, the incidence rate of the primary specific antihypertensive eects did not dier by CKD outcome was 51·6 per 1000 person-years (95% CI status (figure 5; appendix p 43). Similarly, across CKD 50·3–52·9) in the comparator arm versus 45·8 (44·6–47·2) stages, no drug class showed a stage-related gradient of in the treatment arm. Among those without CKD, the eect compared with placebo; estimates were directionally corresponding rates were 30·4 (29·9–30·9) and consistent and of similar magnitude in both early and 26·4 (25·9–26·9), respectively. advanced CKD (figure 5; appendix p 44). Furthermore, we A 5 mm Hg reduction in systolic blood pressure was found no evidence that class-specific eects varied by the associated with a reduced risk of major cardiovascular proteinuria status (figure 5; appendix p 45). disease among participants with CKD (HR 0·91 [95% CI 0·87–0·94]) and participants without CKD (0·90 Discussion [0·88–0·93]), with no evidence of heterogeneity This individual-participant data meta-analysis, pooling (p >0·99; appendix p 37). Similar results were 46 RCTs and comprising 59 185 participants with CKD interaction observed for all secondary outcomes (all p >0·15; and 225 939 without CKD, provides the most interaction appendix p 38). In the analysis stratified by CKD stage, comprehensive randomised evidence on cardiovascular relative treatment eects on risk of major cardiovascular benefits of blood-pressure lowering across the full disease were consistent across all stages, with clear spectrum of CKD, including subgroups with key clinical sustained benefits in advanced CKD stages 4–5 (mean features such as diabetes and proteinuria, and across a eGFR 25 mL/min per 1·73 m²) and no evidence of eect granular range of blood-pressure thresholds at modification (p >0·99; figure 1). For secondary treatment initiation. Each 5 mm Hg reduction in interaction outcomes, although some variation in eect sizes was systolic blood pressure was associated with a lowering observed among subgroups, there was no strong of the risk of major cardiovascular events, regardless of statistical evidence of eect modification, suggesting CKD status. This beneficial treatment eect extended these variations were likely due to chance (all across the full spectrum of CKD severity, including p >0·74; figure 2). Likewise, analyses stratified by advanced CKD stages 4–5 (mean eGFR 25 mL/min per interaction baseline blood pressure showed consistent treatment 1·73 m²), across categories of baseline blood pressure eects across systolic and diastolic blood-pressure categories for either the primary (figure 3) or secondary outcomes, in participants with or without CKD 0·4 (appendix pp 39–40). Baseline proteinuria status did not modify relative treatment eects, suggesting similar benefits in CKD patients with and without proteinuria (figure 4; appendix 0·3 p 41). In contrast, diabetes status modified treatment eects, with attenuated relative risk reductions for major cardiov ascular events in patients with CKD and 0·2 coexisting diabetes compared with those without (p =0·044; figure 4). No such eect modification was interaction observed for secondary outcomes (appendix p 42). We performed several post-hoc analyses. A comple- 0·1 mentary analysis with four subgroups based on CKD and proteinuria status yielded results similar to the main analysis (appendix p 26). Sensitivity analyses—including Fine–Gray competing risk models (appendix p 27), a two-stage random-eects individual-participant data meta- 0 1 2 3 4 5 6 analysis (appendix p 28), uns tandardised treatment eects (appendix p 29), and exclusion of head-to-head trials Figure 1: Kaplan–Meier curves for major cardiovascular disease, by treatment allocation and CKD stage (appendix p 30)—supported the robustness of our primary Cumulative incidence curves for major cardiovascular events stratified by CKD stage, with solid lines representing findings. Absolute treatment eects were also estimated the treatment arm and dashed lines representing the comparator arm. Number-at-risk data are given in the (appendix p 31). appendix (p 25). HRs with 95% CIs were standardised to a 5 mm Hg reduction in systolic blood pressure, estimated from one-stage stratified Cox proportional hazards models. CKD stages were defined with the CKD Epidemiology The network meta-analysis stratified by CKD status and Collaboration 2021 race-free equations:18 stage 1 (eGFR ≥90 mL/min per 1·73 m²), stage 2 (60–89 mL/min per stage included 29 trials: 16 placebo-controlled 1·73 m²), stage 3a (45–59 mL/min per 1·73 m²), stage 3b (30–44 mL/min per 1·73 m²), and stages 4–5 (n=71 399 participants) and 13 head-to-head com parisons (<30 mL/min per 1·73 m²). Major cardiovascular events were defined as fatal or non-fatal stroke or other (n=108 782). For the analysis stratified by proteinuria, cerebrovascular disease, fatal or non-fatal ischaemic heart disease, or heart failure leading to death or hospitalisation. The increasing cumulative incidence with advancing CKD stage reflects the higher baseline 16 trials with available data were included: ten placebo- cardiovascular risk among patients with more severe CKD. Relative treatment benefit was consistent across all CKD controlled (n=9048) and six head-to-head comparisons stages (p >0·99). CKD=chronic kidney disease. eGFR=estimated glomerular filtration rate. HR=hazard ratio. interaction stneve ralucsavoidrac rojam fo ytilibaborp evitalumuC HR (95% CI) CKD stage 1 0·92 (0·87−0·96) CKD stage 2 0·90 (0·87−0·92) CKD stage 3a 0·91 (0·87−0·96) CKD stage 3b 0·89 (0·83−0·96) CKD stage 4−5 0·85 (0·72−1·00) Comparator Treatment Follow-up (years) Articles Mean eGFR Intervention Comparator HR (95% CI) p interaction Events Total Events Total Major cardiovascular events >0·99 Stage 1 99 2791 31 625 3547 34 869 0·92 (0·87–0·96) Stage 2 75 8362 74 107 10 974 84 531 0·90 (0·87–0·92) Stage 3a 54 3306 21 068 4188 23 834 0·91 (0·87–0·96) Stage 3b 39 1198 5643 1512 6350 0·89 (0·83–0·96) Stage 4–5 25 273 1015 322 1092 0·85 (0·72–1·00) Stroke >0·99 Stage 1 99 974 31 647 1174 34 874 0·90 (0·82–0·98) Stage 2 75 2948 74 172 3922 84 585 0·85 (0·81–0·89) Stage 3a 54 1198 21 092 1512 23 869 0·91 (0·84–0·98) Stage 3b 39 394 5646 497 6357 0·88 (0·77–1·01) Stage 4–5 25 73 1015 103 1093 0·75 (0·56–1·02) Ischaemic heart disease >0·99 Stage 1 99 1375 31 651 1834 34 884 0·94 (0·87–1·01) Stage 2 75 4180 74 161 5544 84 574 0·92 (0·88–0·96) Stage 3a 54 1516 21 085 1963 23 860 0·93 (0·87–1·00) Stage 3b 39 499 5649 667 6358 0·89 (0·79–1·01) Stage 4–5 25 109 1017 130 1094 0·87 (0·68–1·13) Heart failure >0·99 Stage 1 99 392 27 349 573 30 606 0·86 (0·76–0·99) Stage 2 75 1303 62 260 1866 72 765 0·87 (0·80–0·94) Stage 3a 54 720 18 554 985 21 518 0·83 (0·75–0·93) Stage 3b 39 343 5174 465 5946 0·86 (0·74–0·99) Stage 4–5 25 106 982 122 1048 0·88 (0·67–1·16) Cardiovascular death >0·99 Stage 1 99 635 31 220 739 34 607 1·01 (0·91–1·13) Stage 2 75 2132 73 479 2778 84 074 0·92 (0·87–0·97) Stage 3a 54 1039 20 984 1376 23 775 0·90 (0·83–0·97) Stage 3b 39 493 5631 589 6346 0·99 (0·88–1·12) Stage 4–5 25 124 1012 139 1094 0·90 (0·70–1·15) All-cause death 0·74 Stage 1 99 1740 31 653 2080 34 877 0·98 (0·92–1·05) Stage 2 75 5617 74 175 6876 84 582 0·98 (0·94–1·01) Stage 3a 54 2536 21 093 3238 23 868 0·92 (0·87–0·97) Stage 3b 39 1125 5647 1329 6359 1·03 (0·95–1·11) Stage 4–5 25 297 1016 359 1095 0·89 (0·76–1·04) 0·50 1·0 2·0 Favours intervention Favours comparator HR per 5 mm Hg reduction in systolic blood pressure Figure 2: Effects of blood-pressure-lowering treatment on primary and secondary outcomes by CKD stage The forest plot shows HRs and 95% CIs per 5 mm Hg reduction in systolic blood pressure, separately for each outcome. HRs and 95% CIs were standardised to a 5 mm Hg reduction in systolic blood pressure, estimated from one-stage stratified Cox proportional hazards models. CKD stages were defined with the CKD Epidemiology Collaboration 2021 race-free equations:18 stage 1 (eGFR ≥90 mL/min per 1·73 m²), stage 2 (60–89 mL/min per 1·73 m²), stage 3a (45–59 mL/min per 1·73 m²), stage 3b (30–44 mL/min per 1·73 m²), and stages 4–5 (<30 mL/min per 1·73 m²). Mean eGFR values within each stage represent the baseline eGFR of participants classified in that subgroup. p values for interaction were derived from likelihood ratio tests comparing models with and without treatment-by-CKD stage interaction terms, assessing heterogeneity of treatment effect across the five CKD stages, and were adjusted for multiple testing with Hommel’s method. Events denotes the number of participants who had the outcome; total denotes the total number at risk. The size of each square is proportional to the inverse variance of the log HR. The vertical line indicates an HR of 1·0 (ie, no effect). CKD=chronic kidney disease. HR=hazard ratio. down to less than 120/70 mm Hg, and irrespective of Few meta-analyses have examined the eect of blood- proteinuria status. Notably, patients with CKD and pressure lowering on major cardiovascular disease and coexisting diabetes derived substantially smaller relative death in patients with CKD, and the available evidence has benefits from blood-pressure lowering than those yielded conflicting results. A Cochrane meta-analysis without diabetes, highlighting a high-risk CKD incorporating six trials comparing more-intensive versus subgroup that might warrant optimised cardiovascular less-intensive blood-pressure targets (n=7348) found no risk management. treatment eect on total cardiovascular disease 1632 Articles Mean SBP (mm Hg) Intervention Comparator HR (95% CI) p interaction Events Total Events Total CKD >0·99 <120 mm Hg 112 184 1146 246 1401 0·88 (0·72–1·08) 120–129 mm Hg 124 333 1999 453 2523 0·93 (0·80–1·07) 130–139 mm Hg 134 621 3476 717 3885 0·96 (0·86–1·07) 140–149 mm Hg 144 785 4422 999 5098 0·91 (0·83–1·01) 150–159 mm Hg 154 766 4255 991 5007 0·95 (0·85–1·05) 160–169 mm Hg 164 710 4939 939 5278 0·83 (0·75–0·92) ≥170 mm Hg 184 1375 7468 1673 8067 0·90 (0·84–0·97) No CKD >0·99 <120 mm Hg 112 506 5123 719 5715 0·80 (0·71–0·90) 120–129 mm Hg 124 967 8913 1277 10 499 0·92 (0·85–1·00) 130–139 mm Hg 134 1545 13 984 2005 16 478 0·94 (0·88–1·00) 140–149 mm Hg 144 2075 19 058 2753 219 80 0·91 (0·85–0·97) 150–159 mm Hg 154 1844 17 432 2527 20 149 0·87 (0·82–0·93) 160–169 mm Hg 164 1737 186 75 2117 19 877 0·90 (0·85–0·96) ≥170 mm Hg 182 2472 22 497 3118 246 46 0·91 (0·86–0·96) 0·50 1·0 2·0 Favours intervention Favours comparator HR per 5 mm Hg reduction in systolic blood pressure Figure 3: Effects of blood-pressure-lowering treatment on major cardiovascular events by baseline CKD status and SBP The forest plot shows HRs and 95% CIs for major cardiovascular events by baseline SBP categories, separately for participants with and without CKD. HRs and 95% CIs were standardised to a 5 mm Hg reduction in systolic blood pressure, estimated from one-stage stratified Cox proportional hazards models. Mean SBP values within each category represent the baseline SBP of participants classified in that subgroup. p values for interaction were derived from likelihood ratio tests comparing models with and without treatment-by-baseline SBP category interaction terms, assessing heterogeneity of treatment effect across the seven SBP categories within each CKD stratum, and were adjusted for multiple testing with Hommel’s method. Events denotes the number of participants who had the outcome; total denotes the total number at risk. The size of each square is proportional to the inverse variance of the log HR. The vertical line indicates an HR of 1·0 (ie, no effect). CKD=chronic kidney disease. HR=hazard ratio. SBP=systolic blood pressure. (relative risk 1·00 [95% CI 0·87–1·15]), cardiovascular cause death (0·90 [0·70–1·16]), and all-cause death (0·90 Intervention Comparator HR (95% CI) p interaction [0·76–1·06]).41 These null findings persisted in analyses Events Total Events Total stratified by only two eGFR categories (<30 mL/min vs Proteinuria status >0·99 30–60 mL/min). An earlier meta-analysis of 11 trials Proteinuria 584 2962 618 3005 0·90 (0·79–1·02) comparing intensive versus standard blood-pressure No proteinuria 1309 11 022 1652 12 081 0·87 (0·80–0.93) targets (n=9287) reported concordant results, showing no Diabetes status 0·044 benefit for cardiovascular disease and no eect on Previous diabetes 1934 8455 2290 9545 0·96 (0·90–1·02) all-cause death.42 In contrast, a meta-analysis using No previous diabetes 2842 19 268 3731 21 726 0·88 (0·84–0·93) broader eligibility criteria with respect to trial design and 0·50 1·0 2·0 intervention (18 trials; n=60 178) found that a 10 mm Hg reduction in systolic blood pressure was associated with a Favours intervention Favours comparator reduction in the risk of major cardiovascular disease HR per 5 mm Hg reduction in systolic blood pressure among patients with CKD.13 Moreover, this study also Figure 4: Effects of blood-pressure-lowering treatment on major cardiovascular disease in people with CKD, identified significant eect modification by baseline CKD stratified by baseline diabetes and proteinuria status status, with a more pronounced relative risk reduction in The forest plot shows HRs and 95% CIs for major cardiovascular events within participants with CKD, stratified by HR observed in the non-CKD group.13 Our study— proteinuria and diabetes status. HRs and 95% CIs were standardised to a 5 mm Hg reduction in systolic blood pressure, estimated from one-stage stratified Cox proportional hazards models. Proteinuria was defined as urine constituting the largest analysis of trial data to albumin-to-creatinine ratio of ≥30 mg/g or urine protein-to-creatinine ratio of ≥0·22 or dipstick ≥1. Previous date—addressed the uncertainties inherent in individual diabetes was defined as a history of diabetes at baseline. p values for interaction were derived from likelihood ratio RCTs and previous meta-analyses that did not have tests comparing models with and without treatment-by-subgroup interaction terms, assessing heterogeneity of sucient statistical power and generalisability across the treatment effect, and were adjusted for multiple testing with Hommel’s method. Sample sizes for proteinuria analysis are smaller because proteinuria data were available only in a subset of trials. Events denotes the number of CKD spectrum. In contrast to previous investigations, we participants who had the outcome; total denotes the total number at risk. The size of each square is proportional to stratified treatment eects by granular CKD stage while the inverse variance of the log HR. The vertical line indicates an HR of 1·0 (ie, no effect). CKD=chronic kidney simultaneously comparing individuals with and without disease. eGFR=estimated glomerular filtration rate. HR=hazard ratio. CKD, and we incorporated comprehensive subgroup analyses by blood-pressure threshold, proteinuria, and diabetes status. Our study fills a crucial evidence gap Articles concerning the ecacy of blood-pressure-lowering Clinical guidelines’ recommendations for blood- therapy for cardiovascular risk reduction in patients with pressure management in CKD vary considerably.7,43–45 The CKD. 2021 Kidney Disease: Improving Global Outcomes A B OR (95% CI) p OR (95% CI) p interaction interaction ACE inhibitor vs placebo 0·69 ACE inhibitor vs placebo 0·64 CKD 0·81 (0·74–0·90) Proteinuria 0·84 (0·65−1·09) No CKD 0·83 (0·78–0·89) No proteinuria 0·78 (0·66−0·92) ARB vs placebo 0·86 ARB vs placebo 0·39 CKD 0·91 (0·82–1·00) Proteinuria 0·91 (0·72−1·15) No CKD 0·92 (0·86–0·98) No proteinuria 0·80 (0·68−0·96) β blocker vs placebo 0·78 β blocker vs placebo 0·82 CKD 0·96 (0·80–1·16) Proteinuria 0·83 (0·41−1·68) No CKD 0·99 (0·89–1·11) No proteinuria 0·91 (0·65−1·29) CCB vs placebo >0·99 CCB vs placebo >0·17 CKD 0·85 (0·76–0·95) Proteinuria 0·90 (0·70−1·15) No CKD 0·85 (0·79–0·92) No proteinuria 0·72 (0·59−0·87) Diuretic vs placebo >0·99 Diuretic vs placebo >0·48 CKD 0·79 (0·69–0·91) Proteinuria 0·60 (0·29−1·20) No CKD 0·79 (0·72–0·86) No proteinuria 0·79 (0·60−1·06) Figure 5: Class-specific effects 0·5 1·0 2·0 0 0·5 1·0 2·0 of antihypertensive drugs on the risk of major Favours intervention Favours placebo Favours intervention Favours placebo cardiovascular disease, stratified by CKD status, C OR (95% CI) p interaction stage, and proteinuria ACE inhibitor vs placebo 0·57 The forest plots show ORs as Stage 1 0·84 (0·73−0·95) the relative treatment effect and their corresponding Stage 2 0·83 (0·78−0·90) 95% CIs for major Stage 3a 0·86 (0·77−0·97) cardiovascular events stratified Stage 3b 0·68 (0·55−0·84) by CKD status (A), proteinuria Stage 4−5 1·06 (0·61−1·86) existence (B), and CKD ARB vs placebo 0·40 stage (C), comparing each Stage 1 0·93 (0·82−1·05) antihypertensive drug class with placebo, estimated from Stage 2 0·91 (0·85−0·98) a Bayesian network meta- Stage 3a 0·94 (0·84−1·06) analysis with fixed-effects Stage 3b 0·80 (0·66−0·97) models. p values for Stage 4−5 0·85 (0·58−1·23) interaction were derived from β blocker vs placebo 0·60 meta-regression comparing Stage 1 0·77 (0·60−0·99) treatment effects across subgroups. The size of each Stage 2 1·05 (0·93−1·19) square is proportional to the Stage 3a 1·01 (0·82−1·24) inverse variance of the log OR. Stage 3b 0·90 (0·59−1·37) The vertical line indicates an Stage 4−5 0·69 (0·16−2·95) OR of 1·0 (ie, no effect). CKD CCB vs placebo >0·79 stages were defined with the Stage 1 0·78 (0·66−0·93) CKD Epidemiology Collaboration 2021 race-free Stage 2 0·88 (0·80−0·95) equations:18 stage 1 (eGFR Stage 3a 0·88 (0·77−1·01) ≥90 mL/min per 1·73 m²), Stage 3b 0·76 (0·60−0·95) stage 2 (60–89 mL/min per Stage 4−5 0·97 (0·62−1·53) 1·73 m²), stage 3a Diuretic vs placebo >0·96 (45–59 mL/min per 1·73 m²), Stage 1 0·79 (0·65−0·94) stage 3b (30–44 mL/min per 1·73 m²), and stages 4–5 Stage 2 0·78 (0·71−0·87) (<30 mL/min per 1·73 m²). Stage 3a 0·81 (0·69−0·96) ACE=angiotensin-converting Stage 3b 0·75 (0·57−0·99) enzyme. ARB=angiotensin Stage 4−5 0·78 (0·34−1·76) receptor blocker. CCB=calcium channel blocker. CKD=chronic 0 0·5 1·0 2·0 3·0 kidney disease. eGFR=estimated glomerular Favours intervention Favours placebo filtration rate. OR=odds ratio. 1634 Articles (KDIGO) guidelines7 recommend a systolic blood- class in CKD mirrors that observed in individuals with pressure target of <120 mm Hg for non-dialysis CKD. preserved renal function. Consequently, class-specific Although this target represents the most intensive evidence of cardiovascular protection from broader recommendation among major guidelines, it is assigned populations can be extrapolated to patients with CKD, a weak strength of recommendation (grade 2B), reflecting including those at more severe stages. Next, hypertension an evidence base largely derived from a single RCT (the in CKD is typically driven by the convergence of multiple SPRINT trial) with a predefined CKD subgroup.7 In biological pathways, including sodium retention, neuro- contrast, the European Society of Hypertension adopts a hormonal activation, endothelial dysfunction, and reduced more conservative primary target of below 140/90 mm Hg, large-artery compliance. Consequently, mono therapy with a with consideration of below 130/80 mm Hg if well single antihypertensive class is often inadequate. Optimal tolerated,46 and the American College of Cardiology– risk management might require a combination therapy American Heart Association recommends below utilising agents from dierent classes. Our findings arm 130/80 mm Hg, encouraging systolic values below that such multiclass regimens can be recommended 120 mm Hg.45 Our stratified analyses provide evidence to flexibly, with consistent cardiovascular ecacy across CKD inform this debate: relative treatment benefits for major severity and proteinuria values. cardiovascular disease are consistent across baseline Although these findings show that blood-pressure blood-pressure categories, extending to values below reduction confers cardiovascular protection across the 120/70 mm Hg. These findings support the more spectrum of CKD and baseline blood pressure, they intensive KDIGO recommendations and strengthen the should not drive an indiscriminate approach to initiating evidentiary foundation for lower blood-pressure targets in therapy in every clinical setting. From a clinical patients with CKD than for those without. perspective, these findings suggest that single baseline Current guidelines uniformly acknowledge the scarcity characteristics (eg, CKD stage, blood pressure, and of randomised evidence for blood-pressure management proteinuria) are not determinants of proportional in patients with more advanced stages of CKD.7,43–46 Our benefit. Instead, clinicians should anticipate consistent study fills this gap and shows consistent relative relative risk reductions across all CKD stages, treatment eects across CKD stages 1–5. Notably, our proteinuria values, and blood-pressure strata, similar to analysis included 14 148 participants with CKD stage 3b those observed in the broader population. Initiating or higher, of whom 2107 had stage 4–5 disease with a treatment necessitates a multifactorial evaluation that mean baseline eGFR of 25 mL/min per 1·73 m². weighs absolute cardiovascular risk against the Likewise, our finding of attenuated relative benefit in likelihood of adverse events, such as acute kidney injury, patients with CKD and coexisting diabetes addresses hyperkalaemia, and symptomatic hypotension. This another explicitly acknowledged evidence gap. KDIGO decision-making process must balance preventive notes that cardiovascular benefits of intensive blood- ecacy with safety while considering comorbidities, pressure lowering cannot be excluded in diabetic CKD, polypharmacy, and patient values. Furthermore, our but remain uncertain.7 Our subgroup analysis suggests stratified analyses examined each clinical factor in that in individuals with both CKD and diabetes, blood- isolation to provide clear evidence across the full pressure lowering oers little or no relative benefit. To spectrum of each characteristic. Real-world decisions mitigate risk in this subgroup, therapeutic regimens require a multidimensional risk–benefit assessment to might require the integration of antihypertensives with establish whether the cardiovascular protection agents such as SGLT2 inhibitors or GLP-1 receptor conferred by blood-pressure-lowering therapy outweighs agonists, which provide robust cardiorenal protection the potential harms for a given patient. and enhance glycaemic regulation.47–49 For SGLT2 Several limitations should be considered when inhibitors specifically, these benefits are preserved in interpreting and generalising the findings of this study. patients with low eGFR despite reduced glucose-lowering We evaluated only relative treatment eects on ecacy.50,51 Given their distinct mechanisms from cardiovascular outcomes; treatment-related adverse antihypertensive drugs, additive or adjunctive eects on events and kidney-specific outcomes were not examined cardiovascular risk seem plausible, especially in CKD because their scope and methodological requirements with diabetes, thus dedicated trials are warranted. dier from those of the present analysis. In the context Renin-angiotensin system (RAS) inhibitors are widely of the benefit–harm balance, concerns have been raised recommended as the cornerstone of antihypertensive particularly in people with advanced CKD. However, therapy for patients with CKD,7,43–45 driven primarily by the trials published in 2025 conducted exclusively in this proven renoprotective eects of RAS inhibitors rather than population did not show an excess risk of serious adverse by definitive evidence of cardiovascular protection in this events,11,12 although evidence from a larger dataset is still population. Our network meta-analysis provides direct needed. A new round of BPLTTC data acquisition, evidence of this gap and carries several key implications for focused on adverse events, benefit–harm evaluation, and clinical practice. First, the relative risk reduction in cost-eectiveness, is underway and will provide cardiovascular disease conferred by each antihypertensive comprehensive evidence to address these endpoints and Articles refine the overall benefit–harm profile of blood-pressure- FS/PhD/21/29110, and FS/PhD/25/29632), the EU (101080430), Roche lowering treatment. Furthermore, we stratified analyses (R94776/CN002), and the Novo Nordisk Oxford Big Data Partnership; royalties or licences from Lucem Health (personal and institutional); by individual clinical features, which is often insucient personal fees from Radclie Cardiology for speaking; personal fees as for identifying distinct patient groups.52 Further research Editor-in-Chief of Heart; and participation on a Medtronic Advisory is warranted to explore potential heterogeneity of Board for Renal Denervation (institutional). DC has received support treatment eects using other phenotypes, including from the UK Research and Innovation Medical Research Council (MR/Y010825/1), the Dunhill Medical Trust (ARVHF2402/7), and the novel approaches to multivariable and high-dimensional National Institute for Health and Care Research (NIR203982) outside of participant stratification.53 For example, compared with the submitted work (the views expressed are not necessarily those of the widely recommended treat-all policy for patients these funders); and received an honorarium as Specialty Chief Editor of with diabetes, a novel AI-based approach to treatment Frontiers in Cardiovascular Medicine (Cardiovascular Epidemiology and Prevention). ZB has received a doctoral fellowship from the British selection successfully deselected 24·3% of individuals, Heart Foundation (FS/PhD/25/29632). MN is supported by an individual with only a very small proportion of false negatives research fellowship from the British Heart Foundation (grant (0·2% of the cohort).54 Data supporting their clinical FS/IPBSRF/22/27060); has received reimbursement and honoraria from utility in CKD warrant further study. AstraZeneca, Nemysis, and Albus Health outside the submitted work; and is the statistical adviser of Heart (BMJ Publishing Group). MW has This meta-analysis, drawing on the largest body of received personal fees from Amgen, Kyowa Kirin, and Freeline. SEK has randomised evidence to date, carries direct and received lecture honoraria from Emcure, Getz, Hikma, JB Pharma, actionable implications for the management of blood Merck, Vector-Intas, and Zydus; and support from the Research Council pressure in individuals with CKD, specifically for the of Norway (grant number 273563). JC has received grants from the National Health and Medical Research Council of Australia. GZ was prevention of major cardiovascular outcomes rather funded by the China Scholarship Council. WCC has received consulting than kidney outcomes or renoprotection. Clinicians fees from Alnylam, Idorsia, and Azurity Pharmaceuticals, and a grant should recommend blood-pressure-lowering treatment from George Medicines. All other authors declare no competing for patients at any stage of CKD and at any baseline interests. blood pressure because it consistently reduces Data sharing cardiovascular risk, irrespective of CKD stage or The governance of the BPLTTC has been reported previously.15 The study is governed by the University of Oxford’s (Oxford, UK) policies on baseline blood pressure. Nonetheless, treatment research integrity and codes of practice, and follows the university’s decisions should also consider the balance of benefits policy on the management of research data and records. The Steering and potential risks, including adverse eects and Committee oversees scientific activities based on BPLTTC datasets. All patient-specific factors. Our results also indicate no trial data shared with the BPLTTC is considered confidential and will not be provided to any third party. Requests for data should be made directly evidence that the eects of antihypertensive drug classes to the data custodians for each trial, whose contact details are available dier across the investigated CKD subgroups, providing through the original trial publications. Information about individual clinicians with the flexibility to select agents based on projects is posted on the BPLTTC website (https://www.wrh.ox.ac.uk/ patient characteristics, preferences, and tolerability. research/Blood_Pressure_Lowering_Treatment_Trialists_Collaboration_ BPLTTC). To ensure transparency and facilitate reproducibility, all Notably, the attenuated treatment eect observed in statistical source code and documentation are deposited in an open- participants with coexisting diabetes underscores the access GitHub repository (https://github.com/deepmedicine/BPLTTC- need for an adopted risk-management strategy in this CKD-at-baseline-and-spectrum) hosted by the DeepMedicine research group (https://www.wrh.ox.ac.uk/research/deep-medicine). subgroup of patients with high-risk CKD. Although blood-pressure lowering remains imperative in patients Acknowledgments The third cycle of the BPLTTC was funded by the British Heart with CKD and diabetes due to their higher absolute risk, Foundation (PG/18/65/33872). This study also received support via an these findings highlight the necessity of considering individual research fellowship from the British Heart Foundation (grant antihypertensive therapy with other evidence-based FS/IPBSRF/22/27060). This study was prepared with research materials interventions to maximise cardiovascular risk reduction from the Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) and the Prevention of Events with in this specific subgroup. Angiotensin-Converting Enzyme Inhibition (PEACE) trial, which were Contributors obtained from the National Heart, Lung, and Blood Institute (NHLBI) GZ, MN, and KR conceptualised the study. GZ, MN, KR, ZB, and DC were Biologic Specimen and Data Repository Information Coordinating responsible for data acquisition, harmonisation, and curation. All authors, Centre. The views expressed do not necessarily reflect the opinions or as members of the working group, were responsible for writing the views of the ALLHAT, PEACE, or the NHLBI. We acknowledge the protocol and conducting the investigation. GZ, MN, ZB, and KR directly original depositors of the Australian National Blood Pressure Study data accessed data. MN and ZB verified data, analytical codes, and results. and the Australian Data Archive and confirm that those who performed All authors interpreted the data. GZ and MN drafted the original the original analysis and data collection bear no responsibility for the manuscript, which was reviewed and edited by all authors. GZ and MN further analysis or interpretation of these data. This study utilised data were responsible for data visualisation. KR, DC, and MN acquired the from a trial supported by Boehringer Ingelheim. Boehringer Ingelheim funding for the study. MN and KR supervised the project. All authors had no role in the design, analysis, or interpretation of the results; (including members of the Blood Pressure Lowering Treatment Trialists’ however, they were given the opportunity to review the manuscript for Collaboration Core Analytic group [ZB, MN, and DC]) had full access to all medical and scientific accuracy regarding Boehringer Ingelheim the data in the study and had final responsibility for the decision to submit substances and for intellectual property considerations. During the for publication. preparation of this work we used Grammarly, integrated into Microsoft Word (Pro, version 1.2.246.1869), to check grammar, punctuation, and Declaration of interests clarity. Claude (Anthropic; Opus 4.5) and ChatGPT Edu (OpenAI; GPT-5) KR reports grants to his institution from the National Institute for were used for debugging R code and structuring the GitHub repository. Health Research (NIHR304997), the Medical Research Council After using these tools, we reviewed and edited the content as needed (MR/Y030419/1), the British Heart Foundation (FS/PhD/22/29321, 1636 Articles and take full responsibility for the content of the publication. The 16 WHO–International Society of Hypertension Blood Pressure authors followed the University of Oxford guidance on the safe and Lowering Treatment Trialists’ Collaboration. Protocol for responsible use of generative AI tools. prospective collaborative overviews of major randomized trials of blood-pressure-lowering treatments. J Hypertens 1998; 16: 127–37. References 17 Canoy D, Copland E, Nazarzadeh M, et al, and the Blood Pressure 1 Rahimi K, Bidel Z, Nazarzadeh M, et al, and the Blood Pressure Lowering Treatment Trialists’ Collaboration. Antihypertensive drug Lowering Treatment Trialists’ Collaboration. Pharmacological blood eects on long-term blood pressure: an individual-level data meta- pressure lowering for primary and secondary prevention of analysis of randomised clinical trials. Heart 2022; 108: 1281–89. cardiovascular disease across dierent levels of blood pressure: 18 Miller WG, Kaufman HW, Levey AS, et al. National Kidney an individual participant-level data meta-analysis. Lancet 2021; Foundation Laboratory Engagement Working Group 397: 1625–36. recommendations for implementing the CKD-EPI 2021 race-free 2 Rahimi K, Bidel Z, Nazarzadeh M, et al, and the Blood Pressure equations for estimated glomerular filtration rate: practical Lowering Treatment Trialists’ Collaboration. Age-stratified and guidance for clinical laboratories. Clin Chem 2022; 68: 511–20. blood-pressure-stratified eects of blood-pressure-lowering 19 Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work pharmacotherapy for the prevention of cardiovascular disease and Group. KDIGO 2024 clinical practice guideline for the evaluation death: an individual participant-level data meta-analysis. Lancet and management of chronic kidney disease. Kidney Int 2024; 2021; 398: 1053–64. 105: S117–314. 3 Nazarzadeh M, Bidel Z, Canoy D, et al, and the Blood Pressure 20 Sarnak MJ, Levey AS, Schoolwerth AC, et al. Kidney disease as a risk Lowering Treatment Trialists’ Collaboration. Blood pressure- factor for development of cardiovascular disease: a statement from the lowering treatment for prevention of major cardiovascular diseases American Heart Association Councils on Kidney in Cardiovascular in people with and without type 2 diabetes: an individual Disease, High Blood Pressure Research, Clinical Cardiology, and participant-level data meta-analysis. Lancet Diabetes Endocrinol 2022; Epidemiology and Prevention. Circulation 2003; 108: 2154–69. 10: 645–54. 21 Appel LJ, Wright JT Jr, Greene T, et al, and the AASK Collaborative 4 Li J, An J, Huang M, et al. Representation of real-world adults with Research Group. Intensive blood-pressure control in hypertensive chronic kidney disease in clinical trials supporting blood pressure chronic kidney disease. N Engl J Med 2010; 363: 918–29. treatment targets. J Am Heart Assoc 2024; 13: e031742. 22 Inker LA, Astor BC, Fox CH, et al. KDOQI US commentary on the 5 Ishida JH, Chauhan C, Gillespie B, et al. Understanding and 2012 KDIGO clinical practice guideline for the evaluation and overcoming the challenges related to cardiovascular trials involving management of CKD. Am J Kidney Dis 2014; 63: 713–35. patients with kidney disease. Clin J Am Soc Nephrol 2021; 16: 1435–44. 23 Nazarzadeh M, Bidel Z, Canoy D, et al, and the Blood Pressure Lowering Treatment Trialists’ Collaboration. Blood pressure 6 Colombijn JMT, Idema DL, van Beem S, et al. Representation of lowering and risk of new-onset type 2 diabetes: an individual patients with chronic kidney disease in clinical trials of participant data meta-analysis. Lancet 2021; 398: 1803–10. cardiovascular disease medications: a systematic review. JAMA Netw Open 2024; 7: e240427. 24 Bidel Z, Nazarzadeh M, Canoy D, et al, and the Blood Pressure Lowering Treatment Trialists’ Collaboration. Sex-specific eects of 7 Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work blood pressure lowering pharmacotherapy for the prevention of Group. KDIGO 2021 clinical practice guideline for the cardiovascular disease: an individual participant-level data meta- management of blood pressure in chronic kidney disease. analysis. Hypertension 2023; 80: 2293–302. Kidney Int 2021; 99: S1–87. 25 Tudur Smith C, Williamson PR. A comparison of methods for fixed 8 Norris K, Bourgoigne J, Gassman J, et al, and the AASK Study eects meta-analysis of individual patient data with time to event Group. Cardiovascular outcomes in the African American Study of outcomes. Clin Trials 2007; 4: 621–30. Kidney Disease and Hypertension (AASK) trial. Am J Kidney Dis 2006; 48: 739–51. 26 Nazarzadeh M, Canoy D, Bidel Z, et al, and the The Blood Pressure Lowering Treatment Trialists’ Collaboration. Methodological 9 Beddhu S, Rocco MV, Toto R, et al, and the SPRINT Research clarifications of recent reports. J Hypertens 2022; 40: 847–52. Group. Eects of intensive systolic blood pressure control on kidney and cardiovascular outcomes in persons without kidney disease: 27 Hommel G. A stagewise rejective multiple test procedure based on a secondary analysis of a randomized trial. Ann Intern Med 2017; a modified Bonferroni test. Biometrika 1988; 75: 383–86. 167: 375–83. 28 European Medicines Agency. Guideline on the investigation of 10 Cheung AK, Rahman M, Reboussin DM, et al, and the SPRINT subgroups in confirmatory clinical trials. Jan 31, 2019. https://www. Research Group. Eects of intensive BP control in CKD. ema.europa.eu/en/documents/scientific-guideline/guideline- J Am Soc Nephrol 2017; 28: 2812–23. investigation-subgroups-confirmatory-clinical-trials_en.pdf (accessed Dec 15, 2025). 11 Rossignol P, Zannad F, Massy Z, et al, and the ALCHEMIST study group. Spironolactone in patients on chronic haemodialysis at high 29 Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in risk of adverse cardiovascular outcomes (ALCHEMIST): medicine—reporting of subgroup analyses in clinical trials. a multicentre, double-blind, randomised, placebo-controlled trial N Engl J Med 2007; 357: 2189–94. and updated meta-analysis. Lancet 2025; 406: 705–18. 30 Law MR, Morris JK, Wald NJ. Use of blood pressure lowering drugs 12 Walsh M, Collister D, Gallagher M, et al, and the ACHIEVE in the prevention of cardiovascular disease: meta-analysis of Investigators. Spironolactone versus placebo in patients 147 randomised trials in the context of expectations from undergoing maintenance dialysis (ACHIEVE): an international, prospective epidemiological studies. BMJ 2009; 338: b1665. parallel-group, randomised controlled trial. Lancet 2025; 31 Wang N, Salam A, Pant R, et al. Blood pressure-lowering ecacy of 406: 695–704. antihypertensive drugs and their combinations: a systematic review 13 Ettehad D, Emdin CA, Kiran A, et al. Blood pressure lowering for and meta-analysis of randomised, double-blind, placebo-controlled prevention of cardiovascular disease and death: a systematic review trials. Lancet 2025; 406: 915–25. and meta-analysis. Lancet 2016; 387: 957–67. 32 van Valkenhoef G, Lu G, de Brock B, Hillege H, Ades AE, 14 Ninomiya T, Perkovic V, Turnbull F, et al, and the Blood Pressure Welton NJ. Automating network meta-analysis. Res Synth Methods Lowering Treatment Trialists’ Collaboration. Blood pressure 2012; 3: 285–99. lowering and major cardiovascular events in people with and 33 Altman DG, Bland JM. Interaction revisited: the dierence between without chronic kidney disease: meta-analysis of randomised two estimates. BMJ 2003; 326: 219. controlled trials. BMJ 2013; 347: f5680. 34 Thompson SG, Higgins JP. How should meta-regression analyses 15 Rahimi K, Canoy D, Nazarzadeh M, et al, and the Blood Pressure be undertaken and interpreted? Stat Med 2002; 21: 1559–73. Lowering Treatment Trialists’ Collaboration. Investigating the 35 Suzuki H, Kanno Y, and the Ecacy of Candesartan on Outcome in stratified ecacy and safety of pharmacological blood pressure- Saitama Trial (E-COST) Group. Eects of candesartan on lowering: an overall protocol for individual patient-level data meta- cardiovascular outcomes in Japanese hypertensive patients. analyses of over 300 000 randomised participants in the new phase Hypertens Res 2005; 28: 307–14. of the Blood Pressure Lowering Treatment Trialists’ Collaboration 36 The Australian therapeutic trial in mild hypertension. Report by the (BPLTTC). BMJ Open 2019; 9: e028698. Management Committee. Lancet 1980; 1: 1261–67. Articles 37 Black HR, Elliott WJ, Grandits G, et al, and the CONVINCE 46 Kreutz R, Brunström M, Burnier M, et al. 2024 European Society of Research Group. Principal results of the Controlled Onset Hypertension clinical practice guidelines for the management of Verapamil Investigation of Cardiovascular End Points (CONVINCE) arterial hypertension. Eur J Intern Med 2024; 126: 1–15. trial. JAMA 2003; 289: 2073–82. 47 Baker WL, Buckley LF, Kelly MS, et al. Eects of sodium-glucose 38 Pepine CJ, Handberg EM, Cooper-DeHo RM, et al, and the cotransporter 2 inhibitors on 24-hour ambulatory blood pressure: INVEST Investigators. A calcium antagonist vs a non-calcium a systematic review and meta-analysis. J Am Heart Assoc 2017; antagonist hypertension treatment strategy for patients with 6: e005686. coronary artery disease. The International Verapamil-Trandolapril 48 Ferdinand KC, White WB, Calhoun DA, et al. Eects of the once- Study (INVEST): a randomized controlled trial. JAMA 2003; weekly glucagon-like peptide-1 receptor agonist dulaglutide on 290: 2805–16. ambulatory blood pressure and heart rate in patients with type 2 39 Schrader J, Lüders S, Kulschewski A, et al, and the MOSES Study diabetes mellitus. Hypertension 2014; 64: 731–37. Group. Morbidity and mortality after stroke, eprosartan compared 49 Kennedy C, Hayes P, Cicero AFG, et al. Semaglutide and blood with nitrendipine for secondary prevention: principal results of a pressure: an individual patient data meta-analysis. Eur Heart J 2024; prospective randomized controlled study (MOSES). Stroke 2005; 45: 4124–34. 36: 1218–26. 50 Heerspink HJL, Stefánsson BV, Correa-Rotter R, et al, and the 40 Braunwald E, Domanski MJ, Fowler SE, et al, and the PEACE Trial DAPA-CKD Trial Committees and Investigators. Dapagliflozin in Investigators. Angiotensin-converting-enzyme inhibition in stable patients with chronic kidney disease. N Engl J Med 2020; coronary artery disease. N Engl J Med 2004; 351: 2058–68. 383: 1436–46. 41 Erviti J, Saiz LC, Leache L, et al. Blood pressure targets for 51 Herrington WG, Staplin N, Wanner C, et al, and the The EMPA- hypertension in people with chronic renal disease. KIDNEY Collaborative Group. Empagliflozin in patients with Cochrane Database Syst Rev 2024; 10: CD008564. chronic kidney disease. N Engl J Med 2023; 388: 117–27. 42 Lv J, Ehteshami P, Sarnak MJ, et al. Eects of intensive blood 52 Kent DM, Steyerberg E, van Klaveren D. Personalized evidence pressure lowering on the progression of chronic kidney disease: based medicine: predictive approaches to heterogeneous treatment a systematic review and meta-analysis. CMAJ 2013; 185: 949–57. eects. BMJ 2018; 363: k4245. 43 NICE. Hypertension in adults: diagnosis and management. UK 53 Li Y, Rao S, Solares JRA, et al. BEHRT: transformer for electronic National Institute for Health and Care Excellence, 2019. health records. Sci Rep 2020; 10: 7155. 44 McEvoy JW, McCarthy CP, Bruno RM, et al, and the ESC Scientific 54 Rao S, Li Y, Mamouei M, et al. Refined selection of individuals for Document Group. 2024 ESC Guidelines for the management of preventive cardiovascular disease treatment with a transformer- elevated blood pressure and hypertension. Eur Heart J 2024; based risk model. Lancet Digit Health 2025; 7: 100873. 45: 3912–4018. 45 Jones DW, Ferdinand KC, Taler SJ, et al. 2025 AHA/ACC/AANP/ AAPA/ABC/ACCP/ACPM/AGS/AMA/ASPC/NMA/PCNA/SGIM guideline for the prevention, detection, evaluation and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2025; 152: e114–218. 1638 --- [PDF原文](https://sci-net.xyz/storage/7932541/7180a61953b34f90d25fbfb1c8990bd8009933379dbbc9edc2cbf623359ed4b1/Pharmacological-blood-pressure-lowering-for-the-prevention-of-cardiovascular-disease-and.pdf) DOI: 10.1016/S0140-6736(26)00367-3