Sleep-time Blood Pressure and Risk of CKD Progression

Hypertension is very common in patients with chronic kidney disease (CKD); its prevalence increases with diminished estimated glomerular filtration rate (eGFR), reaching an estimated 86% in patients with end-stage kidney disease (ESKD). On the other hand, hypertension leads to kidney and other target-organ damage, through its burden of mechanical and oxidative stress on vascular walls.

As in the past, current hypertension guidelines continue to recommend wake-time office blood pressure (BP) measurement (OBPM) as the primary mode of diagnosing hypertension and establishing therapeutic goals. Nonetheless, many of them now advocate ambulatory BP (ABP) monitoring (ABPM) of adult patients to confirm the OBPM-based diagnosis of hypertension because of the well-documented significantly better value of ABPM-derived parameters relative to wake-time OBPM in prognosticating cardiovascular disease (CVD) risk, a relevant finding also well documented in patients with CKD. Unfortunately, ABPM is seldom applied in clinical practice, and when it is, there is no consensus as of yet exactly how to properly apply it, e.g., how frequent to sample BP and for how long, and also what parameter(s) to use for making accurate diagnosis. Most guidelines propose around-the-clock ABPM to derive for diagnostic purpose either the 24h or the "daytime" systolic (SBP) and diastolic BP (DBP) means that typically are defined according to fixed clock time durations established by default by the manufacturers of the measuring devices or set a priori by investigators as opposed to biologically meaningful ones based on the actual clock times of the beginning and end of the activity and sleep spans of each patient to derive accurate individualized awake and asleep BP means and dipping pattern.

Contrary to the recommendation of the most recently published hypertension guidelines to rely on the "daytime" or 24h ABP means to diagnose hypertension when ABPM is performed, multiple prospective outcome trials and meta-analyses substantiate CVD events are much better predicted by the asleep BP mean. CVD risk is additionally predicted by an attenuated sleep-time relative SBP decline - non-dipper (sleep-time relative SBP decline <10%) or riser (sleep-time relative SBP decline <0%) 24h SBP profile. Thus, the elevated asleep SBP mean and blunted sleep-time relative SBP decline (non-dipping) constitute joint significant CVD risk factors, independent of the wake-time OBPM or the awake or 24h ABP means. The importance of the asleep SBP mean for making the diagnosis and prognosis of CVD risk is exemplified by a meta-analysis of the original databases of nine cohorts representing in total 13,844 hypertensive patients. It found that the wake-time office SBP as well as ABPM-derived awake and asleep SBP means were all significantly associated with elevated CVD risk when each variable was analyzed individually; however, when all three SBP measurements were simultaneously included into the survival model, only the asleep SBP mean remained as the independent BP-derived predictor of CVD events.

The differential relevance of several ABPM-derived parameters, compared with the wake-time OBPM, as potential risk markers of CVD morbidity and mortality has been further assessed in the thus far largest reported primary care-based ABPM-based investigation, the Hygia Project, established in 2007 as a multicenter research network - currently comprised of 40 primary care facilities and 292 clinical investigators - designed to incorporate ABPM as routine procedure to diagnose and manage hypertension, assess efficacy of BP-lowering treatment, and evaluate patient CVD and other medical risks. Between 2008 and 2018, participating primary-care physicians - properly trained and certified in the proper application of ABPM and the conduct of procedures of the investigative protocol - referred in total 21,963 primary care patients for 48h ABPM annually, or more frequently when the ABP of treated hypertensive patients remained uncontrolled, i.e., ≥135/85 or ≥120/70 mmHg for, respectively, the awake and asleep SBP/DBP means and those having compelling clinical conditions of elevated CVD risk, including diabetes, CKD, and past major CVD event. During the median follow-up of 6.3 years, 1,830 individuals experienced the main CVD-outcome of CVD death, myocardial infarction, coronary revascularization, heart failure, ischemic stroke, or hemorrhagic stroke. Corroborating and extending previously reported findings - based upon an initial analysis of the data of the Hygia Project cohort of 18,078 individuals that had been recruited up to 2015 - Cox proportional-hazard analyses revealed the asleep SBP mean to be the most significant single BP marker of CVD risk, independent of absence/presence of hypertension therapy at baseline, upon-waking vs. at-bedtime treatment-time, diagnosis of influential morbidities of diabetes or CKD, age, and sex. The joint contribution with the asleep SBP mean to CVD risk was significant only for diminished sleep-time relative SBP decline but not for the wake-time OBPM or the awake or 24h ABP means, such that at any given asleep SBP level, non-dipper individuals showed significantly greater CVD risk than did dipper ones.

A blunted sleep-time BP decline, which is characteristic of the non-dipper/riser 24h BP pattern, is common in patients with CKD. Nonetheless, the reported prevalence of sleep-time hypertension and non-dipping in CKD in different studies is highly inconsistent and, thus, its exact prevalence and associated potential clinical relevance are uncertain. Among other factors, this might be due to differences in the studied populations (treated or untreated patients at differing stages of CKD), relatively small sample sizes, definition of awake and asleep periods by arbitrary fixed clock-hour spans, and frequent reliance only on a single, low-reproducible, 24h ABPM evaluation per participant. Moreover, most previous investigations have evaluated the 24h BP pattern of patients with CKD exclusive of proper comparison to those without CKD.

Several prospective studies have investigated the prognostic value of the features of the 24h BP pattern determined by around-the-clock ABPM for prediction of the development and progression of CKD. Some of these studies, in particular, have found sleep-time hypertension and the non-dipper/rising BP pattern to be predictors of the development and progression of albuminuria, worsening proteinuria, decline in eGFR, and progression to ESKD.

The MAPEC Study investigated the prognostic value of OBPM and ABP to predict new-onset CKD and whether risk reduction might be associated with progressive treatment-induced decrease of wake-time OBPM or ABPM-determined awake or asleep BP means. The authors prospectively evaluated 2,763 individuals without CKD and with baseline ABP ranging from normotension to hypertension. Upon recruitment and annually (or more frequently if hypertension treatment was adjusted based on ABP targets) thereafter, BP and physical activity (wrist actigraphy) were simultaneously monitored for 48h to accurately derive individualized awake and asleep BP means. During a 5.9-year median follow-up, 404 participants developed CKD. CKD risk was greater with progressively higher asleep SBP mean and lower sleep-time relative SBP decline, i.e., more non-dipper BP pattern. The asleep SBP mean was indeed the most significant predictor of CKD in a Cox proportional-hazard model adjusted for age, diabetes, serum creatinine, urinary albumin, previous CVD event, and hypertension treatment-time, i.e., upon-awakening vs. at-bedtime (per 1-SD elevation, hazard ratio 1.44, [95%CI: 1.31-1.56], P<0.001). Wake-time OBPM and awake or 48h ABP means lost their predictive value when corrected by the asleep SBP mean. Analyses of BP changes during follow-up revealed 27% reduction in the risk of CKD per 1-SD decrease in asleep SBP mean, independent of changes in wake-time OBPM or awake ABP mean. In conclusion, sleep-time SBP is a highly significant independent prognostic marker for CKD. Alteration of sleep-time SBP regulation, i.e., increase in asleep SBP mean or blunted sleep-time relative SBP decline, seems to precede, rather than follow, the development of CKD. Furthermore, the progressive treatment-induced decrease of asleep SBP mean, a potential therapeutic target requiring around-the-clock ABPM patient evaluation, might be a significant method for establishing an effective therapeutic regimen to reduce or avert the risk for CKD. Despite these collective findings from ABPM-based clinical trials, current models for predicting the development and progression of CKD either do not include BP as a relevant covariate or rely solely on the wake-time OBPM.

On the other hand, home BP monitoring (HBPM) has been advocated as the preferred method for out-of-office BP assessment by current guidelines to corroborate the diagnosis of hypertension based on wake-time OBPM. This preference seems to be based on some apparent advantages over ABPM, including HBPM being widely used, cost-efficient, and well accepted by hypertensive patients for long-term BP evaluation and potential improvement of compliance with hypertension treatment. The major shortcoming of conventional HBPM devices of the past is that BP measurement could not be programmed and, accordingly, they were not capable of measuring BP during sleep.

Recent advances in self-monitoring device technology have enabled measurement of BP during sleep. The currently available sleep-time HBPM devices provide automatic triggering of a few asleep BP measurements. In particular, the NightView (Omron Healthcare, HEM9601T-E3) is a light (110 g) wrist type HBPM device that, in addition to customary wake-time BP measurements (triggered upon demand by the patient), it also automatically takes BP measurements during sleep. The NightView device has been clinically validated for use in the sitting and in supine position (with palm placed sideways, upwards, and downwards) according to the ANSI/AAMI/ISO81060-2:2013 protocol and it passed all criteria.

The primary objectives of the study are:

1. To prospectively evaluate the prognostic value of sleep-time BP assessed by HBPM and ABPM, as well as its treatment-induced changes during follow-up, as potential makers for progression of CKD.
2. To prospectively evaluate the prognostic value of sleep-time BP assessed by HBPM and ABPM, as well as its treatment-induced changes during follow-up, as potential makers for total and CVD mortality in patients with advanced CKD.
3. To prospectively evaluate the prognostic value of sleep-time BP assessed by HBPM and ABPM, as well as its treatment-induced changes during follow-up, as potential makers for CVD morbidity in patients with advanced CKD.
4. To evaluate, for all of the primary objectives listed above, the potential influence of other HBPM and ABPM-determined parameters of interest as well as their treatment-induced changes during follow-up, including (but not limited to) the sleep-time relative SBP decline (measure of 24h SBP dipping), 48h and awake BP means, measurements of BP variability, ambulatory arterial stiffness index (AASI), morning BP surge, pre-awakening BP surge, and sleep-time BP fall.
5. After assessment of the prognostic value of all potential significant covariates, to develop a risk score system for individualized stratification of patients with advanced CKD.

Required sample size calculation:

If each patient would be evaluated by 48h ABPM only twice, i.e., at baseline and after follow-up, we will obtain two values of each potential prognostic BP variable, but only one of the change in those variables during follow-up. Under this hypothetical situation, at the two-sided a level of 5% and with a power of 80%, 450 participants providing all required information would make possible the detection of a difference of >30% in the incidence of the composite outcome variable after a median follow-up of >1.5 years in patients divided according to each tested prognostic factor or its change during follow-up (e.g., normal/controlled vs. elevated sleep-time SBP mean, dipper vs. non-dipper, etc.). However, by protocol design, each participant will be scheduled for multiple 48h ABPM assessment, i.e., at baseline and then every 3 months during follow up. For a median follow-up of 1.5 years from baseline, participants might provide up to 7 ABPM profiles, with the corresponding 6 values of changes in BP parameters from baseline to each of the quarterly 48h ABPM assessments. Accordingly, due to the unique design of this protocol, at the two-sided a level of 5% and with a power of 80%, a maximum of 160 participants would make possible the detection of a difference of >30% in the incidence of the composite outcome variable after a median follow-up of >1.5 years in patients divided according to each tested prognostic factor or its change during follow-up. Recruitment of an additional up to 20% of the target sample (thus 192 patients in total) will be sought to cover possible dropouts and/or lack of the required 1-year minimum follow-up per participant. These sample size requirements might need to be recalculated as a function of the actual incidence of events (for the defined composite renal+CVD primary outcome) documented during the first year of follow-up.