Blood pressure targets in patients with chronic kidney disease: A critical evaluation of clinical‐trial evidence and guideline recommendations

Hypertension among people with chronic kidney disease (CKD) is an established risk factor for acceleration of atherosclerotic cardiovascular disease and progression of kidney injury to end-stage renal disease (ESRD).1, 2 Randomized trials have demonstrated that blood pressure (BP)-lowering is translated into a profound benefit on cardiovascular and kidney outcomes.1 However, the optimal levels at which BP should be targeted remains an area surrounded by substantial controversy that is reflected by the discrepancy in BP target recommendations for patients with CKD across guidelines released during 2012-2018 (Table 1). The 2012 Kidney Disease Improving Global Outcomes (KDIGO) guideline recommended a lower BP target of 3g/d. In the proteinuria stratum of 0.22 g/d, initial randomization to the lower BP target was associated with 27% reduction in the composite kidney outcome (HR: 0.73; 95% CI: 0.58-0.93). This benefit was not evident in those with baseline proteinuria ≤0.22 g/d (HR: 1.18; 95% CI: 0.93-1.50).11 Another post hoc analysis incorporating data from 840 MDRD participants with an extended follow-up of 9.2 years (trial plus cohort phase) showed that compared with the standard BP target, those initially randomized to intensive BP-lowering had 32% reduced risk of developing ESRD (HR: 0.68; 95% CI: 0.57-0.82).12 Based on this evidence, the 2012 KDIGO guideline recommended a lower BP target of 1 g/d who would presumably carry a higher risk of kidney injury progression were excluded from SPRINT. Although this trial was not originally designed to assess the efficacy of intensive BP-lowering on kidney outcomes like MDRD and AASK, the results of SPRINT were interpreted as a proof of the cardioprotective benefit of this strategy and provided the scientific basis to recommend a lower BP target of <130/80 mmHg for patients with CKD in the 2017 AHA/ACC guideline.4 The same clinical-trial evidence was interpreted differentially by the 2018 ESH/ESC guideline3 and the question that arises is whether the more conservative approach of European hypertension specialists is reasonable and evidence-based. Of note, the 2018 ESH/ESC guideline3 provided a more conservative recommendation on BP targets particularly for patients with CKD and not for other patient populations. Whether this differentiation is attributable to issues related to the tolerability of intensive BP-lowering in the CKD setting (ie, higher prosperity of these patients to adverse events, higher risk of acute kidney injury episodes, etch) or to other reasons remains another area of uncertainty. The diagnostic accuracy, reproducibility, and predictive value of BP recordings taken at the environment of clinic should be interpreted within the context of the actual methodology implemented (such as, type of BP monitor, the seated rest period before BP measurement, the presence of observer, the number of BP recordings).13 In SPRINT, clinic BP was measured with a fully automated oscillometric device that obtained triplicate recordings after a 5-minute seated rest and often without the presence of observer.6 This research-grade BP monitoring technique differs substantially from routine clinic BP recordings. A meta-analysis of 31 diagnostic-test studies (incorporating data from 9279 participants) showed that fully automated clinic SBP was similar with the reference-standard ambulatory daytime SBP mean difference (MD): 0.3 mmHg; 95% CI: −1.1 to 1.7 mmHg.14 In contrast, routine clinic SBP overestimated ambulatory daytime SBP by 14.5 mmHg (MD: 14.5 mmHg; 95% CI: 11.8-17.2 mmHg).14 The importance of standardization in BP measurement methodology was highlighted by another meta-analysis of 10 diagnostic-test studies showing that unattended recordings were similar with attended clinic BP recordings, when the same device and measurement methodology was implemented.15 These meta-analyses, however, quantified only the average differences between BP monitoring techniques and did not explore their actual levels of agreement. In a diagnostic-test study that included 275 patients with stage 3-4 CKD and clinic BP <140/90 mmHg, participants underwent clinic BP monitoring with the research-grade technique that was implemented in SPRINT. On the same day, participants had their clinic BP recorded without specification of a 5-minute seated rest.16 Research-grade clinic SBP was by 12.7 mmHg lower than routine clinic SBP (MD: −12.7 mmHg; 95% CI: −14.7 to −10.7 mmHg) and the 95% levels of agreement between these two techniques were ranging from −46.1 to 20.7 mmHg. Research-grade clinic SBP underestimated the reference-standard ambulatory daytime SBP by 7.9 mmHg (MD: −7.9 mmHg; 95% CI: −9.4 to −6.4 mmHg), whereas the 95% levels of agreement were once again wide and were ranging from −33.2 to 17.4 mmHg. Whereas fully automated clinic SBP and ambulatory daytime SBP were significantly associated with echocardiographically documented left ventricular hypertrophy, routine clinic SBP could not detect evidence of target-organ damage.16 Taken together, the above evidence from diagnostic-test studies suggests that the differentiation in recommended BP targets between the American and European guidelines may be simply the tip of the iceberg. Guideline groups have taken into consideration the above-described variability between research-grade and routine clinic BP recordings. Recognizing that the adoption of a research-grade BP monitoring technique like that of SPRINT in daily clinical practice would probably be problematic, the recommended BP targets were adjusted to a higher level from that of the achieved clinic SBP of 121.2 mmHg in the intensive-arm of SPRINT. In our interpretation, a lower or a higher algebraic adjustment for the average bias inserted by routine clinic BP recordings remains an oversimplification.17 An algebraic adjustment for the average differences provides little to no reflection of the actual variability between research-grade and routine clinic BP at the level of individual patients, given the wide 95% levels of agreement between these two techniques in diagnostic-test studies.16 Thus, implementation of intensive BP targets in daily clinical practice necessitates the optimization of our BP measurement methodology. In other words, if antihypertensive therapy continues to be guided by what is already done in routine clinical practice, then intensive BP-lowering may not be beneficial. Evidence to support that intensive BP-lowering to levels <130/80 mmHg is an effective strategy to delay the progression of kidney injury to ESKD is weak. All randomized trials that compared a lower vs a standard BP target in patients with CKD failed to show a benefit of intensive BP-lowering on kidney outcomes.7-9 The notion that intensive BP-lowering is beneficial for those with proteinuric CKD is based on low-quality evidence from post hoc analysis of the MDRD and AASK with an extended observational follow-up after the completion of the randomized phase of these two trials.11, 12 In accordance with the results of MDRD, AASK, and REIN-2, in a post hoc analysis of 2646 SPRINT participants with eGFR <60 mL/min/1.73 m2, intensive BP-lowering did not improve the composite kidney outcome.10 This post hoc analysis, however, provided some evidence to support a potential benefit of intensive BP-lowering on survival and cardiovascular outcomes.10 Even if we interpret this evidence as a proof of cardioprotection, we believe that this benefit may not be generalizable to the majority of our patients if antihypertensive therapy continues to be guided by BP recordings taken under routine clinical practice conditions. In our view, individualization of the intensity of therapy based on its tolerability and optimization of our BP measurement methodology represent important steps to improve BP control and clinical outcomes in this high-risk population. The wider adoption of home or ambulatory BP monitoring is another important step to improve the management of hypertension among patients with CKD. The authors have no conflicts of interest to disclose. Literature search: PIG, VV; Drafting the manuscript: PIG; Revisions on the initial draft: PEZ, VL; Approval of the final paper: VV, VL, PEZ.