(Editorial comment to: El Jamal N, Brooks TG, Cohen J, et al. Prognostic utility of rhythmic components in 24-h ambulatory blood pressure monitoring for the risk stratification of chronic kidney disease patients with cardiovascular co-morbidity. J Hum Hypertens 38, 420–429 (2024). https://doi.org/10.1038/s41371-023-00884-0).

Chronic kidney disease (CKD) is a growing health issue worldwide; it is linked to increased cardiovascular morbidity and mortality and it is responsible for 1.5% of global deaths [1]. Major cardiovascular events represent the first cause of death in CKD and their incidence progressively increases with advancing CKD stages [2]. The percentage of CKD patients who will develop cardiovascular events or even die from cardiovascular causes before reaching end-stage kidney disease (ESKD) can be as high as 90% [3]. To protect against this increased cardiovascular burden, effective blood pressure (BP) management is of vital importance [4]. Traditional office BP measurements completely miss the BP fluctuations throughout the 24-h interval, a phenomenon more pronounced in patients with CKD. To that purpose, 24-h ambulatory BP monitoring (ABPM) is highly recommended to detect BP variability (BPV), diurnal variations, and nocturnal dipping status [5].

The role of nocturnal dipping and BPV indices calculated by data from 24-h ABPM recordings for cardiovascular risk prediction and stratification in CKD has been previously examined. Higher ambulatory systolic BPV, as assessed via the average real variability (ARV) index, has been associated with increased risk for nonfatal cardiovascular events and all-cause mortality among pre-dialysis CKD patients [6]. Similarly, in ESKD patients undergoing hemodialysis, increased systolic BPV indices during the interdialytic interval, particularly 44-h ARV, have been linked with risk of cardiovascular events and all-cause mortality, independently of the actual ambulatory BP levels [7]. With regards to nocturnal dipping patterns, while non-dipping is recognized as a marker of increased cardiovascular risk in the general hypertensive population [8], the evidence for such an association in CKD patients is weaker. Non-dipping has been associated with left ventricular hypertrophy [9], coronary artery calcification [10] and overall mortality in pre-dialysis CKD [11], but the results were not uniform when cardiovascular morbidity and mortality were evaluated. In several studies, the association of non-dipping status with these outcomes lost its significance after adjustment for daytime BP, 24-h BP or other established cardiovascular risk factors [12, 13]; nonetheless, there is one cohort study in which non-dipping remained a significant predictor of cardiovascular events, even after adjustment for several covariates, including 24-h BP, glomerular filtration rate (GFR), proteinuria and history of cardiovascular disease (CVD) [14].

Predicting cardiovascular events in pre-dialysis CKD based on existing approaches, mainly via dipping status, are inherent to certain limitations. First, these patients display poor reproducibility of their dipping status in consecutive ABPMs, a fact that could partly explain the conflicting findings regarding the association of abnormal dipping with cardiovascular outcomes [15]. Additionally, definitions of nocturnal BP and dipping status have been inconsistent across studies in the general population of CKD patients, with use of different temporal periods for daytime and nighttime intervals [16, 17]. Moreover, dipping status is commonly confounded by sleep/awake timing and does not explore the characteristics of a time series dataset. Given that nocturnal dipping is typically assessed as a binary variable, the complexities of BPV and diurnal variations are oversimplified [18]. These facts necessitate the use of more robust analytical methods.

In a recent issue of the Journal of Human Hypertension, El Jamal et al. [19] investigated a novel approach for detecting diurnal rhythmic components in ABPM in order to better assess and stratify the cardiovascular risk associated with BP patterns. The authors applied the JTK_CYCLE algorithm, a non-parametric method originally designed to characterize oscillations and cyclical variables over a 24-h period in large datasets deriving from gene expression studies. Of interest, the JTK_CYCLE algorithm has the advantage of capturing rhythmic components without requiring accurate sleep onset and offset data, a challenge in many clinical settings.

The study analyzed 24-h ABPM data from two large cohorts of pre-dialysis CKD patients, the Chronic Renal Insufficiency Cohort (CRIC) and the African American Study for Hypertension and Kidney Disease (AASK) [19]. Nocturnal dipping was assessed by calculation of dipping ratio (DR) (i.e., dividing mean nighttime BP over mean daytime BP) and BPV by calculation of AVR and afterwards participants were categorized based on present or absent rhythmic components using the JTK_CYCLE algorithm. Rhythmic components were evident in 34% of CRIC and 26% of AASK participants. Of note, 24-h BP curves from CRIC participants with similar DR and ARV indices were clearly discriminated by rhythmic components quantified by the JTK_CYCLE algorithm. Predictors of rhythmic components and dipping patterns were different and only overlapped in some demographic traits and existing comorbidities. In particular, African American race, diabetes, and history of CVD in the CRIC cohort and obesity (body mass index > 30) in the AASK cohort were significantly associated with the absence of rhythmic components. On the contrary, significant predictors of non-dipping were African American race, age >65 years, uncontrolled BP, history of CVD and protein-creatinine ratio (PCR) > 500 mg/g in the CRIC cohort and uncontrolled BP, diabetes, and use of beta-blockers in the AASK cohort.

The study’s most compelling finding is the prognostic value of rhythmic components of BP. Rhythmic profiling of BP identified subgroups of CRIC participants that were more likely to die from cardiovascular causes. In the fully adjusted model, a trend towards a significant association between absent cyclic components and cardiovascular death in the full CRIC cohort was evident (HR = 1.71, 95% CI: 0.99–2.97, p = 0.056). CRIC participants with prior CVD and absent cyclic components in their BP profile had 3.4-fold higher risk for cardiovascular death compared to participants with prior CVD and cyclic components present in their BP profile (HR = 3.37, 95% CI: 1.45–7.87, p = 0.005). Additionally, the absence of rhythmic components compared to their presence significantly increased the incidence rate of cardiovascular mortality (18.46 [14.30–23.82] compared to 5.04 [2.65–11.22] per 1000 person years respectively, p < 0.001); this finding was neither present for 24-h BP control nor dipping status. In contrast, ARV tertiles were not significantly associated with cardiovascular death in unadjusted and adjusted models. The above results were not replicated in the AASK cohort, perhaps due to large differences in patient characteristics and low number of events in the latter.

This study yields important findings with promising clinical translation. First, the detection of rhythmic components in ABPM by the JTK_CYCLE algorithm may represent a useful prognostic indicator of cardiovascular mortality among CKD patients with prior CVD, an approach that shifts away from the simplistic binary categorization of BP dipping status. This finding combined with the fact that ARV was not linked to cardiovascular death in this study, shows that the JTK_CYCLE algorithm could detect a discrete high-risk population compared to patients characterized so far by abnormal dipping or increased BPV. Further research is needed to validate these findings with prospective studies examining whether this or similar sophisticated tools for ABPM analysis could offer better risk stratification and, thus, mitigation of cardiovascular complications in CKD patients.