Intensive BP Control in Patients with CKD and Risk for Adverse Outcomes

Abstract

Significance Statement

Although most guidelines recommend tightly controlling BP in patients with CKD, individuals with advanced kidney disease or severe albuminuria were not well-represented in trials examining the effect of this intervention on kidney outcomes. To examine the effect of intensive BP control on the risk of kidney outcomes in patients with CKD, the authors pooled individual-level data from seven trials. They found that overall, intensive BP control was associated with a 13% lower, but not significant, risk of a kidney outcome. However, the intervention's effect on the kidney outcome differed depending on baseline eGFR. Data from this pooled analysis suggested a benefit of intensive BP control in delaying KRT onset in patients with stages 4–5 CKD, but not necessarily in those with stage 3 CKD.

Background

The effect of intensive BP lowering (to systolic BP of <120 mm Hg) on the risk of kidney failure requiring KRT remains unclear in patients with advanced CKD. Such patients were not well represented in trials evaluating intensive BP control.

Methods

To examine the effect of intensive BP lowering on KRT risk—or when not possible, trial-defined kidney outcomes—we pooled individual-level data from seven trials that included patients with eGFR<60 ml/min per 1.73 m². We performed prespecified subgroup analyses to evaluate the effect of intensive BP control by baseline albuminuria and eGFR (CKD stages 4–5 versus stage 3).

Results

Of 5823 trial participants, 526 developed the kidney outcome and 382 died. Overall, intensive (versus usual) BP control was associated with a lower risk of kidney outcome and death in unadjusted analyses but these findings did not achieve statistical significance. However, the intervention's effect on the kidney outcome differed depending on baseline eGFR (P interaction=0.05). By intention-to-treat analysis, intensive (versus usual) BP control was associated with a 20% lower risk of the primary kidney outcome in those with CKD GFR stages 4–5, but not in CKD GFR stage 3. There was no interaction between intensive BP control and the severity of albuminuria for kidney outcomes.

Conclusions

Data from this pooled analysis of seven trials suggest a benefit of intensive BP control in delaying KRT onset in patients with stages 4–5 CKD but not necessarily with stage 3 CKD. These findings suggest no evidence of harm from intensive BP control, but also point to the need for future trials of BP targets focused on populations with advanced kidney disease.

Podcast

This article contains a podcast at https://dts.podtrac.com/redirect.mp3/www.asn-online.org/media/podcast/JASN/2023_02_27_JASN0000000000000060.mp3

Introduction

The Kidney Disease Improving Global Outcomes (KDIGO) guidelines recently changed their recommendation to treat all patients with CKD to a target clinic systolic BP (SBP) of <120 mm Hg (level 2B evidence), assuming that clinic BP is measured using a standardized approach.¹ While individualized decision-making was suggested, the guidelines do extend to populations who were not well represented or even excluded in major clinical trials, such as those with advanced CKD. Although the results of Systolic Blood Pressure Intervention Trial (SPRINT) showed benefits from intensive BP control in terms of cardiovascular disease (CVD) and mortality risk in patients with CKD, SPRINT did not demonstrate clear evidence of benefit or harm regarding CKD progression.^2,3 This is likely because of the low number of individuals who developed kidney failure requiring replacement therapy (KFRT) during the relatively short follow-up duration of SPRINT and other major hypertension trials that have included patients with CKD.^3–6

To date, there are an insufficient number of clinical trials designed to test intensive BP lowering exclusively in populations at very elevated risk of progression to KFRT or death, such as those with stage 4 or 5 disease or severe albuminuria, although some high-risk individuals were included in select trials.⁷ Lack of trial-grade evidence to guide their care has required the extrapolation of the benefit or harm of BP-lowering interventions on kidney and mortality outcomes using data from patients with milder forms of disease. The lack of definitive evidence to support the implementation of intensive BP lowering in these high-risk populations could lead to lower uptake of guideline recommendations in routine clinical practice, especially because BP control remains inadequate even in populations with milder forms of CKD who were well-represented in prior trials.⁸ In addition, individuals with advanced CKD are at higher risk for acute kidney injury from excessive BP reduction and the occurrence of hyperkalemia if using renin-angiotensin-aldosterone blockers to achieve lower BP targets,⁹ and scarce trial data exist that inform clinicians as to the balance of safety and efficacy.

The objective of this study was to test the effect of intensive BP control in patients with CKD (defined as eGFR <60 ml/min per 1.73 m²) using pooled data from seven large clinical trials of intensive BP control. In particular, we were interested in whether the effects of intensive BP control had benefits for kidney or mortality outcomes in patients with CKD stage 4 or 5 compared with patients with CKD stage 3, but we also prespecified analyses to examine heterogeneity in the effect of BP control by the severity of albuminuria.

Methods

Data Sources

The data from the trials included in this study were derived from the National Heart, Lung, and Blood Institute BioLINCC Repository, the National Institute of Diabetes and Digestive and Kidney Diseases Central Data Repository, and the National Institute of Neurological Disorders and Stroke Archived Clinical Research Datasets, which provide public access to these deidentified data, or shared directly by trial investigators. The study was reviewed and approved by the University of California San Francisco and Tufts University Institutional Review Board. All trials obtained informed consent from participants at enrollment.

Trial Selection

Trials were only included if they enrolled patients with CKD stage 3 or above and tested interventions that included a component of SBP lowering (targeting SBP or mean arterial pressure [MAP]). Our focus was on trials that included SBP as a target given the recent emphasis on guidelines on the treatment of SBP.¹⁰ The seven trials that we identified for inclusion were the Modification of Diet in Renal Disease (MDRD) Trial, African American Study of Kidney Disease and Hypertension (AASK), Action to Control Cardiovascular Risk in Diabetes (ACCORD), the SPRINT, Secondary Prevention of Small Subcortical Strokes Trial (SPS3), Effect of Strict Blood Pressure Control and Angiotensin Converting Enzyme inhibition on the Progression of Chronic Renal Failure in Pediatric Patients (ESCAPE), and Ramipril in Efficacy in Nephropathy (REIN-2) Trial. We required that trials included for analysis had at least 15 kidney failure events to ensure the inclusion of a high-risk population. Trials that randomly assigned participants to different antihypertensive agents but not different BP targets were excluded from the study. We considered the inclusion of the Strategy of Blood Pressure Intervention in the Elderly Hypertensive Patients (STEP trial) conducted in China that randomized participants to intensive versus less intensive SBP control, but there were an insufficient number of patients of KFRT among the 196 participants with baseline CKD enrolled in the trial.¹¹

Study Population

Only patients with eGFR<60 ml/min per 1.73 m² were included for pooled analysis to examine the effect of intensive BP lowering on the risk of KRT, or when not possible, trial-defined kidney outcomes. In all trials, race was defined according to trial protocol and most commonly by self-report.

MDRD Study was a 2×2 factorial trial⁶ of the effect of intensive BP control and low-protein diet on CKD progression. Conducted from 1989–1993 (N=840), adults 18–70 years with nondiabetic causes of CKD were randomly assigned to an MAP of ≤92–98 mm Hg (∼BP 125/75 mm Hg) versus a MAP of ≤102–107 mm Hg (∼BP 140/90 mm Hg), depending on the age of the participant. AASK^4,12–14 was a 2×3 factorial trial of Black participants with hypertensive CKD (N=1094) assigned to intensive (MAP≤92 mm Hg) versus usual BP control (MAP 102–107 mm Hg) and different BP agents between 1995 and 2001. Participants were 18–70 years with GFR 20–65 ml/min per 1.73 m². Both MDRD and AASK trials found that intensive BP control did not have a statistically significant effect on the primary kidney outcomes (defined as a reduction in kidney function or need for KRT).^6,12

SPRINT² was a trial of intensified BP control among >9000 adults without diabetes, of whom 2646 had CKD at baseline enrollment. Patients ≥50 years at elevated risk for cardiovascular events were randomly assigned to a goal SBP of <120 versus 140 mm Hg. Intensive BP control reduced the risk of a major cardiovascular event^2,3 but did not have a statistically significant effect on the predefined primary kidney outcome (defined as a composite of a 50% decline in kidney function or need for KRT).

ACCORD⁵ was a 2×2 factorial trial of intensified BP and glycemic control among patients with type II diabetes at elevated CVD risk, of whom 499 had CKD. Patients were randomly assigned to a goal SBP of <120 versus 140 mm Hg, but there was not a statistically significant effect on the primary outcome of a major cardiovascular event during the parent trial.⁵

SPS3 trial was a 2×2 factorial design trial of intensive BP control (to SBP<130 mm Hg) versus usual BP control (SBP 130–149 mm Hg) in patients with a prior history of a lacunar stroke, in addition to clopidogrel versus placebo for secondary stroke prevention.¹⁵ During the trial, 405 patients had CKD at baseline enrollment, but the trial did not have a statistically significant effect on its primary outcome, which was the recurrence of a stroke.

REIN-2¹⁶ was a trial of intensive BP control (BP<130/<80 mm Hg) versus usual BP control (diastolic BP<90 mm Hg) in 338 adults with nondiabetic chronic kidney disease. During the trial, intensive BP control did not have a statistically significant effect on delaying the progression to KFRT.

ESCAPE¹⁷ (N=385) was a trial of intensive BP control (to ambulatory MAP<50th percentile) versus usual BP control (ambulatory MAP 50th–95th percentile for age and sex) in children <18 years with CKD. During the trial, intensive BP control was found to be effective in reducing the risk of the primary composite outcome of a 50% decline in eGFR, eGFR<10 ml/min per 1.73 m², or onset of KRT.

Definition of the Trial Intervention

Participants were classified as receiving intensive versus usual BP control on the basis of individual trial definitions.

Outcomes

The primary kidney outcome in this study was the onset of KRT before death during the parent trial unless otherwise specified to ensure consistency in the length of follow-up across different studies (only MDRD, AASK, and ESCAPE trials had long-term follow-up posttrial closure) and ensure maintenance of different BP targets in the randomized groups, which contrasts with other prior studies.^18–20 For the ESCAPE trial, the kidney outcome we used was a composite onset of KFRT or eGFR<10 ml/min per 1.73 m². For ACCORD and SPS3 studies, the trial-defined kidney outcome that we adopted was a composite of either onset of KFRT, eGFR<15 ml/min per 1.73 m², or serum creatinine>3.3 mg/dl at any study visit.

The secondary outcome was death before KFRT as ascertained by each of the individual trials during the parent study. Administrative censoring for all outcomes occurred at the time of trial closure.

Subgroups of Interest

We tested for interaction by baseline eGFR (as a continuous variable) using the Chronic Kidney Disease Epidemiology Collaboration 2009 equation and severity of albuminuria, which were prespecified factors of interest. In children, eGFR was computed using the bedside Schwartz equation.²¹ We then chose to present our subgroup analyses on the basis of thresholds that classify different stages of CKD (e.g., eGFR< versus ≥30 ml/min per 1.73 m²) since these are clinically meaningful stages of the disease. Subgroup analyses were only performed if we detected an interaction (which, given the low power that is common with tests of interactions, we defined as P<0.10) with the randomized BP intervention assignment.

Statistical Analyses

We pooled all trials for analysis as a single trial without consideration of the study recruitment site. For trials that only had measurements of urine protein/creatinine ratio (UPCR), these values were converted to urine albumin/creatinine ratio (UACR) using a validated conversion formula.²² Neither UPCR nor UACR were available in the SPS3 trial data that were archived in the repository and thus UACR was imputed using data from the other trials.²³ Multivariable regression models with 50 imputations were used to impute the missing logarithm of UACR, accounting for age, Black race, eGFR, and sex.

We used the Cox proportional hazards regression model to examine the association between the randomized assignment to the BP intervention and risk of the onset of the primary kidney outcome, and secondarily, death with follow-up censored at the time of KFRT. All models accounted for differences in the baseline hazard for outcomes across the different trials in Cox models. We examined whether there was heterogeneity across trials by testing for an interaction between the randomized BP intervention and trial, and we also analyzed each trial separately to understand the differences in effect sizes across the various trials for the primary kidney outcome.

We considered our primary analyses to be unadjusted, given that the intervention was randomly assigned in all trials. In the secondary analysis, we adjusted models for age, sex, race, history of diabetes, baseline eGFR, and baseline logarithm of albuminuria to control for residual confounding. These covariates were selected on the basis of their association with the progression of CKD and their availability across all trials.

In sensitivity analysis, we also examined the effect of the BP intervention on the risk for the onset of the kidney outcome with the use of Fine-Gray models to account for the competing risk of death. We repeated our primary analysis with the exclusion of ESCAPE trial participants so that only adults were included for analysis of the kidney and death outcome. In addition, we repeated our primary analysis excluding the SPS3 trial, and then excluding both ACCORD and SPS3 trials which had trial-specific kidney outcomes that included events other than KFRT.

Due to the presence of interaction by baseline eGFR (as a continuous variable), we performed additional subgroup analysis by the stage of CKD at baseline (stages 4–5 versus stage 3).

To ensure that any interactions noted in our analyses were not driven by between-trial differences, in additional sensitivity analyses, we determined the mean eGFR for each trial and also centered eGFR for each trial. We then examined the interaction terms between randomized BP arm assignment and the mean eGFR of each trial, and between randomized BP assignment and the difference between the centered mean eGFR and each individual's eGFR in both unadjusted and adjusted analyses. Next, we additionally included interaction terms between each covariate of interest and trial data source and repeated our unadjusted and adjusted analyses.²⁴

Finally, we repeated our analysis on the basis of the achieved SBP between month 6 and the end of follow-up and compared outcomes using unadjusted and adjusted logistic regression to determine whether the findings were similar to our primary analysis. We adjusted for the same covariates as above using only adult trials (given that in children, BP levels need to be percentile by age and an absolute threshold would not be applicable across different age ranges). In these analyses, achieved SBP was categorized as <120 mm Hg in adults, 120–140 mm Hg, and >140 mm Hg.

All statistical analyses were performed using STATA 16 (StataCorp, LLC).

Results

We included 5823 trial participants who had a baseline eGFR<60 ml/min per 1.73 m². The mean age of participants at randomization was 62 years 40% were women, and 31.6% were of Black race. The mean baseline eGFR was 43 ml/min per 1.73 m². Characteristics of individuals were balanced between the two randomized treatment arms (Table 1). Approximately 43% of patients included in the study were from SPRINT, 8.6% were from ACCORD trial, 14.1% from MDRD Study, 16.5% from AASK trial, 7.0% were from SPS3 trial, 5.6% from ESCAPE trial, and 5.3% from REIN-2 Trial. The characteristics of participants in each of the individual trials who were included in our analysis are shown in Supplemental Table 1.

Table 1.

Baseline characteristics of individuals at randomization

Characteristic	Usual BP	Intensive BP	Total
Mean±SD or N (%) unless otherwise noted	N=2891	N=2932	N=5823
Age (yr) at randomization	61.7±18.1	61.6±18.1	61.7±18.1
Women	1157 (40.0)	1196 (40.8)	2353 (40.4)
Black race	924 (32.0)	914 (31.2)	1838 (31.6)
Baseline eGFR (ml/min per 1.73 m2)	42.5±12.7	42.8±12.7	42.7±12.7
Median urine albumin/creatinine ratio (mg/g) with IQRa	29.1 (7.9–231.8)	28.6 (8.1–245.8)	28.7 (8.1–239.7)
SBP (mm Hg)	139.1±19.7	139.1±20.6	139.1±20.1
DBP (mm Hg)	78.6±14.5	79.0±14.6	78.8±14.5
Diabetes	346 (12.0)	362 (12.3)	708 (12.2)

IQR, interquartile range; SBP, systolic BP; DBP, diastolic BP; SPS3, Secondary Prevention of Small Subcortical Strokes.

^aExcludes SPS3 Trial due to missing albuminuria.

A total of 526 patients developed kidney outcomes during 241,266 person-years of follow-up. Overall, the rate of onset of the primary kidney outcome was 0.22 per 100 person-years in the pooled cohort. The event rate varied across the different trials, ranging from 0.02 per 100 person-years in SPRINT to 0.98 per 100 person-years in REIN-2 Trial (Supplemental Table 2).

The effect of intensive (versus usual) BP control on the onset of the primary kidney outcome for each individual trial is shown in Supplemental Figure 1 and Supplemental Table 2. There was no evidence of heterogeneity in the effect of intensive BP lowering across the different trials for the kidney outcome (all P>0.1 for interaction between the BP intervention and trial) or the mortality outcome (all P>0.1 for interaction).

Findings in the Overall Cohort

In unadjusted analysis, intensive BP control reduced the risk of kidney outcome by 13% and also reduced the risk of death by 13% (Figure 1 and Table 2), but the findings did not achieve statistical significance. Findings were similar in adjusted analysis for the primary kidney outcome and all-cause mortality. When we treated death as a competing risk for the kidney outcome, findings were also similar in unadjusted (sub hazard ratio [HR], 0.82; 95% confidence interval [CI] 0.72 to 1.02) and adjusted analyses (sub-HR, 0.84; 95% CI, 0.70 to 1.01).

Figure 1.

Risk of the primary kidney outcome by randomized assignment to the intervention.

Table 2.

Association between intensity of BP control and risk of kidney outcome or death in the overall pooled cohort

Overall Cohort	Kidney Outcome	Death
No. of events	526	382
HR (95% CI)
Unadjusteda	0.87 (0.74 to 1.04)	0.87 (0.71 to 1.06)
Adjustedb	0.86 (0.72 to 1.02)	0.87 (0.71 to 1.06)
Excluding ESCAPE Trial
Unadjusteda	0.89 (0.74 to 1.07)	0.87 (0.71 to 1.06)
Adjustedb	0.87 (0.72 to 1.05)	0.87 (0.71 to 1.06)

HR, hazard ratio; CI, confidence interval.

^aAllows baseline hazard to differ across different cohorts.

^bAdjusted for age, sex, Black race, baseline eGFR, logarithm of urine albumin/creatinine ratio, and diabetes allowing for baseline hazard to differ across different cohorts.

When we excluded the ESCAPE trial to only focus on adults, findings in the remaining 5497 participants were also consistent for the primary kidney outcome in unadjusted analysis (HR, 0.89; 95% CI, 0.74 to 1.07) and adjusted analysis (HR, 0.87; 95% CI, 0.72 to 1.05). For the outcome of death, findings were also unchanged with the exclusion of ESCAPE trial (HR, 0.87; 95% CI, 0.71 to 1.06 in unadjusted and adjusted analyses).

In sensitivity analysis when we additionally excluded SPS3 trial from our analysis after excluding ESCAPE trial (N=5092), the hazard ratio for the risk of KRT was 0.88 (95% CI, 0.73 to 1.06) in unadjusted analysis and 0.87 (95% CI, 0.72 to 1.05) in adjusted analysis and consistent with our primary results. For the outcome of death, after excluding ESCAPE and SPS3 trials, the hazard ratio was also similar to that seen in our primary analysis (unadjusted HR, 0.88; 95% CI, 0.71 to 1.09 and adjusted HR, 0.89; 95% CI, 0.71 to 1.11).

When we additionally excluded ACCORD trial after excluding ESCAPE and SPS3 participants (N=4593 included for analysis), the risk for the kidney outcome was similar to the primary analysis (unadjusted HR, 0.90; 95% CI, 0.74 to 1.09 and adjusted HR, 0.90; 95% CI, 0.74 to 1.10). The risk for death was also similar to the results from our primary analysis (unadjusted HR, 0.84; 95% CI, 0.66 to 1.07 and adjusted HR, 0.86; 95% CI, 0.67 to 1.09).

Subgroup Analysis by CKD Stage

There was an interaction (unadjusted P=0.05; adjusted P=0.09) between randomized assignment to the BP intervention and baseline eGFR (as a continuous variable) for the kidney outcome. Intensive BP control was associated with a 20% lower risk of kidney outcome in stage 4 or 5 CKD, but not in stage 3 CKD (Figure 1 and Table 3). There was no evidence of interaction between assignment to the BP intervention group and baseline eGFR for death (p_interaction=0.55 in unadjusted and P=0.52 in adjusted analyses).

Table 3.

Association between intensity of BP control and risk of KRT by baseline stage of CKD in subgroup analysis

HR (95% CI)	CKD Stage 3	CKD Stage 4 or 5
N	4743	1080
Kidney outcomes
No. of events	372	154
Unadjusted	1.08 (0.79 to 1.48)	0.80 (0.65 to 0.98)
Adjusted	1.00 (0.72 to 1.38)	0.81 (0.65 to 0.998)
Death before onset of need for KRT
No. of events	334	48
Unadjusteda	0.82 (0.66 to 1.02)	1.26 (0.71 to 2.23)
Adjustedb	0.82 (0.66 to 1.01)	1.29 (0.72 to 2.31)

HR, hazard ratio; CI, confidence interval.

^aAllows baseline hazards to differ across different cohorts.

^bAdjusted for age, sex, Black race, baseline eGFR, logarithm of urine albumin/creatinine ratio, and diabetes allowing for baseline hazard to differ across different cohorts.

In sensitivity analysis, when we excluded ESCAPE trial participants, an interaction persisted between randomized assignment to the BP intervention and baseline GFR treated as a continuous variable (p_int=0.02 in unadjusted analysis and p_int=0.04 in adjusted analysis). In unadjusted analysis, the risk of KRT was lower (HR, 0.78; 95% CI, 0.62 to 0.97) in those receiving intensive (versus usual BP control) in those with stage 4–5 CKD, but not in those with stage 3 CKD (HR, 1.20; 95% CI, 0.86 to 1.66). Findings were consistent when we additionally excluded ACCORD and SPS3 trials in unadjusted analysis (HR, 1.23; 95% CI, 0.84 to 1.79 in CKD stage 3 and HR, 0.80; 95% CI, 0.65 to 0.99 for CKD stages 4–5, P=0.049) for the outcome of KRT.

When we performed sensitivity analysis centering our effect modifier (eGFR) and including interaction terms between each trial and covariate of interest, the results were similar. There was an interaction between randomized assignment to the BP intervention and the continuous difference between each individual's eGFR and the mean eGFR for each trial in both unadjusted (P=0.004) and adjusted analysis (P=0.037) when treating eGFR as a continuous effect modifier. When we additionally adjusted for interaction terms between covariates included in the model and trial data source, the P-value for interaction between randomized assignment to the BP intervention and eGFR as a continuous variable (P=0.056 for interaction) remained similar to our primary analysis when we did not center the effect modifier or the covariates (Supplemental Table 3).

There was no evidence of an interaction between randomized assignment to the BP intervention and logarithm of albuminuria in unadjusted (P=0.36) or adjusted analysis (P=0.57) for kidney outcomes. There was also no interaction between the randomized assignment to the BP intervention and logarithm of albuminuria in unadjusted (P=0.67) or adjusted analysis (P=0.68) for the outcome of death. Therefore, additional subgroup or sensitivity analyses were not performed.

Analyses by Achieved BP in the Overall Cohort

When we repeated our analysis using the average achieved SBP level in adult trials as the predictor (N=5157), we found that compared with those who achieved a mean SBP of >140 mm Hg, those who achieved an SBP of <120 mm Hg or an SBP of 120–140 mm Hg had lower odds of developing the kidney outcome (odds ratio [OR] 0.29; 95% CI, 0.23 to 0.37 and OR 0.43; 95% CI, 0.33 to 0.57, respectively) in unadjusted analysis. Findings were slightly attenuated but still statistically significant in adjusted analysis (OR 0.43; 95% CI, 0.33 to 0.57 and OR 0.58; 95% CI, 0.48 to 0.71) for the kidney outcome.

When we repeated our analysis using achieved SBP for the outcome of mortality, we found that compared with those who achieved a mean SBP of >140 mm Hg, those who achieved an SBP of <120 mm Hg or an SBP of 120–140 mm Hg had a lower odds of death (OR 0.76; 95% CI, 0.54 to 1.06 and OR 0.76; 95% CI, 0.58 to 0.99, respectively) in unadjusted analysis. Findings were attenuated in adjusted analysis (OR 0.93; 95% CI, 0.65 to 1.32 and OR 0.85; 95% CI, 0.64 to 1.13).

Discussion

The management of BP in patients with CKD remains an area of significant controversy, and the current recommended BP treatment goal for the population with CKD is inconsistent across different guidelines.^25–30 Data from our pooled cohort analysis suggest that in patients with CKD, intensive BP control reduced the risk of the primary kidney outcome by 13%, although this finding did not reach statistical significance. However, we identified heterogeneity in the effect of intensive BP control on baseline kidney function. Specifically, in patients with advanced CKD stage 4 or 5, there was evidence of a 20% reduction in the risk of progression to the primary kidney outcome in those receiving intensive BP control that was not observed in those with CKD stage 3. We did not observe any effect modification by the severity of albuminuria at baseline enrollment for the kidney outcome or all-cause mortality. There was also a tendency toward reduced risk of death with intensive BP lowering, although this finding was not statistically significant. We did not identify heterogeneity in the effect of intensive BP control target on the outcome of death by baseline kidney function or albuminuria.

The KDIGO BP Workgroup recently updated its guidelines to recommend treating all patients with CKD to a target SBP of <120 mm Hg, assuming BP is measured using standardized procedures in the clinic.¹⁰ This recommendation relies primarily on the results of SPRINT, which showed CVD benefit for the CKD subgroup.³ The STEP trial and post hoc analyses of ACCORD trial have also suggested that intensive BP control was associated with lower CVD risk in those randomly assigned to intensive BP control and less intensive glycemic control,^11,31 whereas other trials have evaluated the effect of intensive BP control on kidney or CVD outcomes and did not find beneficial effects.¹² However, there remains a lack of solid evidence surrounding the risk or benefit of intensive BP lowering in subgroups that are not well represented in individual clinical trials, including those with advanced CKD (stage 4 or 5 disease).³² Fewer than 20% of patients included in our study had stages 4 and 5 CKD (N=1080 across seven different trials), and the short trial duration of most clinical trials limited each individual trial from drawing meaningful conclusions within this high-risk population. The strength of our study included the conduct of an individual-level pooled analysis to examine the effect of intensive BP control on kidney outcomes in patients with advanced CKD across multiple trials, which allowed us to limit event follow-up to the trial duration and enhance power.

To date, most guidelines on BP targets have extrapolated data from patients with a higher level of kidney function to those with advanced CKD. Although we may have expected that patients with CKD stage 4 or 5 would derive lesser benefit from intensive BP control given their higher risk for acute kidney injury and other adverse outcomes, such as hyperkalemia, these individuals were a subgroup that seemed to derive the most benefit from intensive BP lowering. While it is unclear why patients with stages 4 and 5 CKD are more likely to derive kidney benefit from intensive BP control in comparison with stage 3 CKD, we suggest that these findings may be related to the high risk of kidney failure in this population. We recognize that the currently available data do not definitively answer the question as to whether there are differences in the benefit of tight BP by stage of CKD or what the appropriate BP target is for patients with advanced CKD. Our results, however, do highlight the potential equipoise of conducting additional clinical trials to address these questions.

The KDIGO guidelines also extrapolated the recommended SBP target of <120 mm Hg to persons with diabetes given that the CVD benefits of intensive control were similar in ACCORD trial participants who received less intensive BP control to an SBP target of <120 mm Hg with standard glycemic control.³¹ In this pooled analysis of seven major trials that have included patients with CKD, only 12% of the population had a history of diabetes. We believe that the low representation of individuals with diabetes warrants attention during the consideration of the optimal BP targets for kidney protection in the diabetic CKD population because the available trial data may be insufficient to draw definitive conclusions.

The strengths of our study include the large sample size and number of patients whom we were able to include with advanced CKD in an individual-level pooled analysis. We included children to provide a comprehensive overview of the trial data that are available because it relates to intensive BP control, although we recognize that there may be distinct differences in the etiology of kidney disease and the setting in which BP was measured in pediatric versus adult studies (ambulatory BP monitoring versus clinic BP). This was the rationale for repeating our analyses with the exclusion of ESCAPE trial participants. Other limitations of our study included the lack of direct proteinuria measurements in SPS3, which required the use of multiple imputations. In addition, the results of our pooled cohort analysis may be driven by AASK, MDRD, and REIN-2 studies—in whom the highest rate of kidney events occurred. We acknowledge that pooling trials with too few events to address important questions may not improve precision and power to resolve uncertainty, and our confidence intervals and significance test results cannot be regarded as definitive. We also do not have detailed granular data on antihypertensive medication use in all trials. It is possible that trial-level confounding and aggregation bias may be present when pooling data across trials for analysis and testing for effect modification. Finally, trial participants may not be representative of the general population with CKD seen in routine clinical practice.

In conclusion, data from this pooled cohort analysis of seven trials suggest the benefit of intensive BP control in patients with stage 4 or 5 CKD from the perspective of kidney outcomes. Although our analyses are post hoc, there is no evidence of harm in this high-risk population from either the kidney or the mortality standpoint. However, we acknowledge that these data require further confirmation in adequately powered clinical trials. Efforts should be devoted to including such underrepresented populations in clinical trials, and to test interventions in those with advanced CKD (on the basis of eGFR or albuminuria) who are at high risk for progression to KFRT to provide more definitive evidence on the appropriate BP targets in such populations.

Supplementary Material

jasn-34-385-s001.docx ^{(27.1KB, docx)}

Disclosures

B.A. Grimes reports Ownership Interest: Cisco, Mattel, and Oracle. L. Inker reports Consultancy: Diamtrix; Research Funding: funding to institute for research and contracts with the Chinnocks, National Institutes of Health, National Kidney Foundation, Omeros and Reata Pharmaceuticals; consulting agreements with Tricida Inc.; Advisory or Leadership Role: Alport Foundation—Medical Advisory Council, NKF—Scientific Advisory Board; and Other Interests or Relationships: American Society of Nephrology member, National Kidney Foundation member. E. Ku reports Ownership Interest: Edison Company; Research Funding: CareDX, Natera, NIH; and Advisory or Leadership Role: American Kidney Fund Health Equity Coalition, AJKD Associate Editor. G. Remuzzi reports Consultancy: Alexion Pharmaceuticals, AstraZeneca, Menarini Ricerche Spa, Otsuka, Silence Therapeutics; and Advisory or Leadership Role: member of numerous Editorial Boards of Scientific Medical Journals. M.J. Sarnak reports spouse works for Eli Lilly; Consultancy: Akebia (Steering Committee of a trial, Funds paid to Tufts Medical Center, Cardurian Consultant; and Research Funding: NIH). F. Schaefer reports Consultancy: Akebia, Amgen, Alexion, Alnylam, Astellas, AstraZeneca, Bayer, Boehringer Ingelheim, Fresenius Medical Care, GSK, Otsuka, Purespring, Relypsa, Roche; Research Funding: Fresenius Medical Care, Novartis, Roche; Honoraria: Amgen, Kyowa Kirin, Otsuka, Roche; Patents or Royalties: Springer (textbook royalties); and Advisory or Leadership Role: Scientific Advisory Board activities for Alexion, Otsuka. E. Wühl reports Advisory or Leadership Role: Editorial Board Member: Pediatric Nephrology, Journal of Hypertension, Executive Board Member German Hypertension League (Deutsche Hochdruckliga, DHL), Vize-Chair COST Action HyperChildNET (EU Program Horizon 2020); Advisory Board activities: Alnylam Pharmaceuticals (Advisory Board member), Desitin and, Hartmann. All remaining authors have nothing to disclose.

Funding

This study was funded by National Institute of Diabetes and Digestive and Kidney Diseases grant R01DK121904 to E. Ku, C.E. McCulloch, L.A. Inker, H. Tighiouart, and M.J. Sarnak.

Author Contributions

L. Inker, E. Ku, C. McCulloch, M. Sarnak, and H. Tighiouart conceptualized the study; L. Inker, G. Remuzzi, P. Ruggenenti, M. Sarnak, F. Schaefer, H. Tighiouart, and E. Wühl were responsible for data curation; B. Grimes, E. Ku, and C. McCulloch were responsible for formal analysis; E. Ku wrote the original draft; M. Sarnak was responsible for funding acquisition; L. Inker, E. Ku, G. Remuzzi, P. Ruggenenti, M. Sarnak, F. Schaefer, and E. Wühl were responsible for resources; B. Grimes, L. Inker, C. McCulloch, M. Sarnak, and H. Tighiouart provided supervision; B. Grimes, C. McCulloch, and M. Sarnak were responsible for methodology; L. Inker, E. Ku, G. Remuzzi, P. Ruggenenti, M. Sarnak, F. Schaefer, H. Tighiouart, and E. Wühl were responsible for investigation; and B. Grimes, L. Inker, C. McCulloch, G. Remuzzi, P. Ruggenenti, M. Sarnak, F. Schaefer, H. Tighiouart, and E. Wühl reviewed and edited the manuscript.

Data Sharing Statement

The data from the trials included in this study were derived from the National Heart, Lung, and Blood Institute BioLINCC Repository, the National Institute of Diabetes and Digestive and Kidney Diseases Central Data Repository, and the National Institute of Neurological Disorders and Stroke Archived Clinical Research Datasets, which provide public access to these deidentified data, or shared directly by trial investigators.

Data used in this paper cannot be shared because: For REIN-2 and ESCAPE trial, these data were generously shared by parent trial investigators for meta-analysis in this study and could be available on request and with further discussion with the parent trial investigators. However, the main study investigators under our data use agreement do not have the authority to rerelease such data without the permission of the parent trial investigators.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/D639.

Supplemental Table 1. Characteristics of participants in each trial.

Supplemental Table 2. Adjusted risk for kidney and death outcomes by each trial.

Supplemental Table 3. Additional analyses examining interactions between randomized BP assignment and eGFR as a continuous variable.

Supplemental Figure 1. Unadjusted risk of the kidney outcome in each trial.