Day–night blood pressure variations: mechanisms, reproducibility and clinical relevance

One of the great advantages offered by ambulatory blood pressure monitoring (ABPM) is the possibility to describe BP patterns over 24 h under daily life conditions. This has produced a deeper insight into BP physiology, as well as into the association of certain diseases with characteristic circadian BP profiles. Subsequent to the application of the pioneering techniques for intra-arterial beat-by-beat ABPM over 24 h [1], it has become clear that BP is far from being maintained at a fixed level. Instead, BP shows continuous and often marked fluctuations [2] with a variety of time constants, ranging from a few seconds to hours, days and months, as in the case of seasonal BP fluctuations [3,4]. Analysis of 24-h ABPM data has allowed a quantitative description of BP variability. This is the case when considering all BP changes at the same time, summarized by the standard deviation of the 24-h average BP. This is also the case when focusing on specific components contributing to overall BP variance, such as the morning BP rise and the nocturnal BP fall [3,5]. The BP fall occurring at the time of night sleep is among the components of BP variability that have received the greatest attention by both physiologists and clinicians. The many studies investigating the mechanisms underlying these diurnal BP changes have highlighted the role played by several factors, such as central neural influences related to sleep stages, with a major contribution by the autonomic nervous system, humoral factors such as corticosteroid hormones and the renin–angiotensin system, and respiratory factors related to the mechanics of ventilation and the accompanying changes in arterial blood O2 and CO2 concentrations [3,6]. In 1988, O'Brien et al. [7] reported for the first time that an abnormal circadian blood pressure profile with decreased night-time dipping may lead to a higher risk of stroke. Subsequent studies of populations [8–11] and hypertensive cohorts [12–18] provided corroboration that an elevated nocturnal blood pressure is a harbinger of an unfavourable outcome. In spite of the apparent concordance between these large-scale outcome studies [8–18], several potential limitations require further clarification with respect to the prognostic accuracy of daytime versus night-time ambulatory blood pressure. Many studies considered only fatal outcomes [8,9,16,17] or did not have the power to study cause-specific cardiovascular endpoints [8,9,11,15]. Investigators dichotomized the night-to-day blood pressure ratio, and applied different definitions of dipping status or the daytime and night-time intervals. Few reports formally compared the predictive value of the blood pressure at night over and beyond the daytime level. Finally, in predominantly treated cohorts of hypertensive patients, antihypertensive drug treatment attenuated the association between outcome and blood pressure [13]. An additional important problem to consider in this field is the reproducibility of BP changes between day and night. Such a reproducibility and, as a direct consequence, the reproducibility of patients' classification into dippers and nondippers is indeed still a matter of lively debate. In this issue of the Journal, Hernandez-del Rey et al.[19] provide additional data on this topic originating from the National ABPM Registry of the Spanish Society of Hypertension. This was done by obtaining 48-h ABP recordings from 611 hypertensive patients, 235 of whom were yet untreated, recruited throughout Spain. The aim of the study was to investigate the reproducibility of the circadian ABP pattern over a 48-h period, by comparing day–night BP changes observed during the first 24-h period of ABPM with those observed during the second 24-h period, as well as with the mean day–night changes over the entire 48-h recording. The percentage of patients classified as nondipper for the first 24 h, the second 24 h and, on average, over the entire 48-h period was 47, 50 and 48%, respectively. When comparing the first with the second 24-h intervals, 24% of patients switched their classification from dipper to nondipper, or vice versa, with consistent results, when separately considering systolic and diastolic BP or treated and untreated patients. The conclusion drawn by the authors is that the categorization of essential hypertensive patients, investigated in a general practice setting, into dippers or nondippers based on a single 24-h ABPM is only moderately reproducible because one out of five patients changes her/his circadian BP profile classification over the next 24 h. Concerning the question of whether a longer duration of ABPM (e.g. 48 h) might provide more reliable data on day–night BP changes, the authors suggested that the average day–night BP difference derived from two 24-h periods may be more reproducible than that quantified by the analysis of one 24-h period only. A precise demonstration of this, however, is not provided in their study, in which no repetition of 48-h ABP recordings over time is available, whereas the comparison made by the authors between the day–night BP changes during the first 24 h and those averaged over 48 h is methodologically questionable [19]. The reproducibility of circadian BP profiles has been investigated in a number of previous studies (Table 1) [19–30], with differing results that were influenced by a number of factors. These include the methods selected to assess the day–night BP changes, the focus on treated or untreated subjects, the patients' concomitant clinical conditions, and the degree of daytime physical activity [31] or the quality of night-time sleep [24], although controversial data have been published also on this issue [32,33].Table 1: Reproducibility of nocturnal blood pressure fallChaves et al.[28] investigated the reproducibility of the dipping status in 101 subjects, including both normotensive and treated hypertensive patients, in whom 24-h ABPM was performed three times at intervals of 8–15 days. When the dipper/nondipper status was defined based on a predefined cut-off point (nocturnal BP reduction higher or lower than 10% of daytime BP levels, respectively), such a definition was poorly reproducible in the subsequent recordings. Conversely, when the percentage BP decline was analyzed as a continuous rather than as categorical variable, the average nocturnal BP fall was not different between recordings. Chaves et al.[28] did not report comparisons between recordings within individual subjects. Ben-Dov et al.[29] retrospectively examined the reproducibility of BP dipping at night in clinical practice patients with duplicated ABPM by accounting for the sleep–awake states, rather than by relying on arbitrary day–night definitions. They found that systolic BP dipping was equally reproducible as the average 24-h BP, with 66% of patients showing a reproducible dipping status. In their study, awake BP included also BP recorded during night-time arousals, whereas sleep BP also included BP recorded during afternoon naps. In a number of other studies, however, a limited reproducibility was reported for nocturnal BP fall in hypertensive patients. Palatini et al.[20] analyzed data from 508 patients included in the HARVEST trial, comparing BP dipping observed in two ambulatory recordings performed 3 months apart, and concluded that the reproducibility of night-time BP dipping was poor. Mochizuki et al.[22] assessed the day–night BP changes in 253 never-treated hypertensive patients and found a 29% variability in the dipper/nondipper definition between repeated recordings, whereas Manning et al.[24] reported a variability of approximately 50%. James et al.[21] reported that more than one-third of 42 elderly hypertensive patients modified their dipping status over a 2-month period. Omboni et al.[23] found that approximately 40% of hypertensive patients included in SAMPLE (Study on Ambulatory Monitoring of Pressure and Lisinopril Evaluation) changed their dipping pattern when ABPM was repeated after 1 year. Finally, the reproducibility of nocturnal BP dipping can be importantly affected by other diseases in the presence of hypertension. Rahmana et al.[27] investigated 59 hemodialysis patients, in whom the ambulatory BP was monitored for 44 h on three occasions, at baseline and after 6 and 12 months, respectively. The overall reproducibility of the dipper/nondipper classification was poor: 55% of patients maintained their original classification at 6 months and 70% at 12 months. However, when the nondipper status was considered separately, 92% of initially nondipper patients remained nondippers at 6 and at 12 months. The authors concluded that haemodialysis patients identified as nondippers consistently reproduce the same circadian profile over long-term follow-up, which is in agreement with the data provided by Covic et al.[25], but not in line with the results obtained by Peixoto et al.[26]. The latter study showed a reproducibility of nondipping of between 53 and 78% in the same type of patients. Similar results on a high short-term reproducibility of the nondipping pattern have been shown by Cuspidi et al.[30] in 36 hypertensive patients with type 2 diabetes, as compared with 61 untreated nondiabetic hypertensive subjects, by performing 24-h ABPM twice over a 4-week period. Some 91.3% of diabetic hypertensive patients, who were nondippers at the first recording, showed the same circadian pattern during the second ABPM, with the reproducibility of the dipping pattern being 84.6%. Conversely, in nondiabetic hypertensive patients, reproducible dipper and nondipper patterns were observed only in 49.2 and 29.5% of patients, respectively. The practical conclusion of the authors was that a nondipper pattern appears to be reliable in diabetic subjects, even if assessed by a single performance of 24-h ABPM, whereas repeated ABP recordings should be recommended to define the circadian BP pattern in a correct manner in nondiabetic individuals. An international consortium [34] recently reported risk estimates independently associated with the daytime and night-time BP in 7458 subjects (mean age = 56.8 years; 45.8% women), enrolled in prospective population studies in Denmark, Belgium, Japan, Sweden, Uruguay and China. Median follow-up was 9.6 years. Adjusted for daytime BP, night-time BP predicted total (n = 983; P < 0.0001), cardiovascular (n = 387; P < 0.01) and noncardiovascular (n = 560; P < 0.001) mortality. Conversely, adjusted for night-time BP, daytime BP predicted only noncardiovascular mortality (P < 0.05), with lower blood pressure levels being associated with increased risk. Both daytime and night-time BP consistently predicted all cardiovascular events (n = 943; P < 0.05) and stroke (n = 420; P < 0.01). Adjusted for night-time BP, daytime BP lost prognostic significance only for cardiac events (n = 525; P ≥ 0.07). Adjusted for the 24-h BP, the night-to-day BP ratio predicted mortality, but not fatal combined with nonfatal events. Antihypertensive drug treatment removed the significant association between cardiovascular events and the daytime BP. Thus, in the international database [34], the predictive accuracy of the daytime and night-time BP and the night-to-day BP ratio depended on the disease outcome under study and differed for fatal outcomes compared to the composite of fatal and nonfatal diseases [34]. For fatal endpoints, night-time BP performed better than the daytime BP, and the night-to-day BP ratio predicted mortality. By contrast, for fatal combined with nonfatal outcomes, the daytime BP performed equally as well as the night-time BP and the night-to-day BP ratio lost its prognostic accuracy. Antihypertensive drug treatment was a major confounder because patients with more severe hypertension and those with a history of cardiovascular complications were at the highest cardiovascular risk but, at the same time, were also more likely to be treated. Hypertensive patients take their medications during daytime and the blood pressure-lowering activity often weans off at night. This mechanism predictably leads to lower daytime BP, higher night-time BP, and a diminished night-to-day BP ratio. Reverse causality might also contribute to the inconsistency in the prediction of disease outcomes by the night-time as opposed to the daytime BP. The international consortium reported that reverse dippers, who had higher night-time than daytime BP, were at the highest risk of all-cause mortality [34]. In several studies [14,17,35], reverse dippers not only were more frequently on antihypertensive drug treatment [14,17], but also they were older [14,17,35] or more likely to have a history of diabetes mellitus [35] or previous cardiovascular disease [14,35]. Moreover, similar to cardiovascular risk, the night-to-day BP ratio also increases with advancing age [36]. In the international database [34], participants with a systolic night-to-day BP ratio of 1 or more were older and therefore at higher risk of death, but they died at an older age than those whose night-to-day ratio was normal (≥ 0.80 to < 0.90). Thus, reverse dipping might be a marker rather than a cause of a worse outcome. The inverse associations in the international database [34] between noncardiovascular mortality and the daytime BP and between total mortality and the daytime BP in untreated subjects also supported the interpretation of reverse causality. In conclusion, the classification of patients into dippers and nondippers depends heavily on arbitrary criteria, is poorly reproducible, and has a different prognostic meaning according to the disease outcome under study, the prevailing 24-h BP level, and treatment status. We would therefore recommend that, in future studies, any categorical results of the night-to-day BP ratio be supported by continuous analyses adjusted for the 24-h BP and be stratified for treatment status. Moreover, the available evidence supports the concept that the ambulatory BP should be recorded over the whole day, as both the night-time and daytime BP levels carry prognostic information. The 24-h BP level [37], rather than the dipping pattern, should continue to inform clinical decisions.