Figures




Figure 1. Out of 1943 individuals in this study who received a placebo, 48% responded to this "treatment", most of them within less than a year (8). A high percentage of responders to placebo renders it likely that "office hypertension" or "white-coat hypertension" associated with the excitement of one or a few visits to a physician's office leads to the entry into large-scale studies of many false positive cases. A substantial error in diagnosis may well be made, not only at the outset of a study in terms of false positives entering the study, but also at the final evaluation in which the ratio of "cured normotensives" is likely to be complicated by a substantial number of false negative diagnoses. Similar errors in diagnosis are likely to occur in the general practitioner's office when relying on isolated blood pressure readings. © Halberg.



Figure 2. The traditional approach is unsatisfactory in view of the meta-analysis of Wilcox et al. (9) who suggest that clinical practice should not be based on results of such clinical trials. This should be qualified: as such studies are conducted without minimal chronobiologic provisions at a total cost of hundreds of millions of U.S. $. © Halberg.



Figure 3. There is a large overlap between the distributions of MESORs (rhythm-adjusted means) of systolic blood pressure of patients conventionally diagnosed as normotensive or hypertensive (11). False positives and false negatives account for about 43% of all conventional diagnoses. © Halberg. Original data of I. Kuwajima.



Figure 4. Three around-the-clock blood pressure profiles at intervals of a few weeks on presumably normotensive individuals, lasting between 2 and 4 days, indicate that systolic blood pressure excess is found in 11 of 23 patients (48%), but consistently only in three of them (13%). Since deviations may be detected in one profile but not in all three profiles, blood pressure monitoring over minimal spans of 1 week is suggested (12). © Halberg. Original data of F. Raab.


Figure 5a.. Original data of M. Haus. Figure 5b.. Original data of J.C. Menendez, analyzed by A. Portela. Figure 5c. Figure 5d.
Figure 5. Large decreases in blood pressure can occur spontaneously that warrant the cessation of anti-hypertensive medication (41) (a). There are also day-to-day changes in circadian amplitude of blood pressure, observed in the case of a 33-year-old neurosurgeon (b) and of a 63-year-old woman under conditions of restricted activity (c). Some of the trends in blood pressure recur every year, as seen in the case of a 70-year-old psychiatrist who shows a circannual rhythm in his blood pressure with an unusually large amplitude (d). These and other (e.g., Figure 4) results, notably from long-term ambulatory monitoring, underlie the suggestion of a minimal 7-day monitoring span, repeated quarterly for at least one year and, if need be, for even longer-term if not continuous self-monitoring. © Halberg.



Figure 6. Changes in circadian pattern of systolic and diastolic blood pressure and heart rate as a function of age from newborns to centenarians (males). Each curve represents hourly means averaged over all individuals within a given group. Data from the International Womb-to-Tomb Chronome Initiative. © Halberg.



Figure 7. A larger amplitude of blood pressure is found in children with as compared to those without familial antecedents of an elevated blood pressure and related vascular disease. © Halberg.



Figure 8. Reliance on casual blood pressure measurements is undesirable because blood pressure abnormality may occur at odd times of the day or night, when it is unlikely to be checked, as documented here for the diastolic blood pressure of a patient treated once a day in the morning. The automatic, ambulatory monitoring of blood pressure of this patient revealed nightly elevated values in 11 of 12 monitoring spans, each of 24 hours, recognized as excess (blackened area), by comparison of the patient's profile with the time-specified upper 95% prediction limits of peers. Because both the duration and the extent of excess are accounted for, excess is expressed in mm Hg x h; it is calculated herein over consecutive 3-hour spans covering an idealized 24-hour cycle; hence, it is denoted as the "3-hourly fractionated hyperbaric index". Such a diastolic blood pressure excess for several hours around midnight in this diurnally active, nocturnally resting man should prompt concern and added timed treatment. © Halberg.



Figure 9. Relative risk of ischemic stroke and nephropathy in patients with an excessive circadian blood pressure amplitude by comparison to risk of patients with an acceptable circadian blood pressure amplitude. For different classes of the 24-hour mean (MESOR) of systolic blood pressure, even below 130 mm Hg, an excessive circadian blood pressure amplitude is associated with a relative risk larger than 1, that is with an increase in risk of ischemic stroke and nephropathy. This increase in risk is statistically significant for normotensive patients, as illustrated by the fact that the 95% confidence interval does not overlap 1 (equal risk). A relative risk of 3.4 represents an increase in risk by 240% (e.g., for the case of ischemic stroke of patients with a MESOR of systolic blood pressure between 140 and 150 mm Hg). © Halberg. Original data of K. Otsuka (34).



Figure 10. Relative risk of ischemic stroke for various risk factors computed as the ratio of the incidence of ischemic stroke that occurred in patients presenting with the tested risk factor by comparison with that in patients not presenting with the tested risk factor. Results of a 6-year prospective study on 297 patients indicate that the risk associated with an excessive circadian blood pressure amplitude is larger than that of all other risk factors considered (obesity, high cholesterol, male gender, drinking alcohol, having familial antecedents with high blood pressure or adverse vascular event, smoking, age above 60 years, and an elevated mean blood pressure value). As compared to patients with an acceptable circadian blood pressure amplitude, patients with an excessive circadian diastolic blood pressure amplitude have a risk of ischemic stroke 8.2 times larger (i.e., they have an increase in risk of 720% to develop an ischemic stroke). © Halberg. Original data of K. Otsuka (34).



Figure 11. Relative risk of ischemic stroke associated with an excessive circadian blood pressure amplitude. In order to test for any interaction of the risk from an excessive circadian blood pressure amplitude with that from other known risk factors (obesity, high cholesterol, male gender, smoking, consumption of alcohol, familial antecedents, old age and MESOR-hypertension), the relative risk was computed in subpopulations not presenting with the tested risk factor other than the circadian blood pressure amplitude. In each case, an excessive circadian blood pressure amplitude represents a larger risk factor for ischemic stroke than the tested risk factor. The fact that the 95% confidence intervals almost invariably do not overlap 1 indicates that an excessive circadian blood pressure amplitude raises the risk of ischemic stroke statistically significantly, irrespective of the effect of the other risk factor tested. © Halberg. Original data of K. Otsuka (34).



Figure 12. Separate comparisons of results from once-traditional treatment with equal doses three times a day versus time-targetted (chrono-)therapy for propranolol, clonidine, and a-methyldopa: each comparison involves 40 patients, randomly assigned to either traditional treatment or chronotherapy (20 patients per group). Chronotherapy consists of drug administration 1.5 to 2 hours before the circadian peak(s) in systolic blood pressure, determined after synchronization for 3 days of each untreated patient with the hospital routine. The single initial dose selected for chronotherapy consisted of 50 to 70% of the dose used for traditional treatment. With this "handicap for chronotherapy design", patients were matched by stage of disease and, as possible, by age, gender, and clinical signs. Daily morning blood pressure measurements, and a diary by the patient listing any headache, chest pain, dizziness, palpitation, insomnia, or other symptoms helped in assessing the time to the appearance of the desired effect and the incidence of side effects, respectively. The 24-hour profile of systolic blood pressure was repeated after 2 weeks, the dose of the drug being changed in the interim if and as needed. A clinically stable hypotensive effect was detected earlier for chronotherapy as compared with traditional therapy, with, on the average, smaller doses and greater efficacy. Side effects from treatment were reduced in the case of chronotherapy (35). These pioneering results must not be extended to ACE inhibitors and calcium antagonists, drugs with other chronopharmacokinetics. The design in these studies is best replaced by longitudinal monitoring with control charts (Figure 18). © Halberg. Data of R. Zaslavskaya.



Figure 13. A 33-year-old neurosurgeon diagnosed conventionally as moderately hypertensive reveals large variation and the need for dense and long sampling in order to approach the dilemma whether to treat or not to treat. Original data of Dr. J.C. Menendez analyzed by Dr. A. Portela. © Halberg.



Figure 14. The conventional interpretation of the data collected during office hours by the 33-year-old neurosurgeon referred to in Figure 13 leaves one in a quandary if only 56% of the measurements are acceptable but 44% are found to be elevated over a 23-day monitoring span. © Halberg.



Figure 15. The decision to treat or not to treat the 33-year-old neurosurgeon referred to in Figures 13 and 14 is the more difficult since during office hours most of his measurements (77%) are acceptable on one day but are invariably too high on another day. © Halberg.



Figure 16. Control charts of daily mean values of blood pressure and heart rate computed on all data collected at 15-minute intervals are shown with a shaded decision interval. While the series of daily means is proceeding "in control" (i.e., at the pre-treatment mean level), the cusum comprises two line graphs that generally stay within the "decision interval" limits, which are plotted here as the horizontal lines at 4.4 and -4.4. Two curves signal increase and decrease in mean, respectively. The breaking by the dashed curve out of the decision interval boundary provides the rigorous validation of the decrease in daily blood pressure mean (by 1 standard deviation). The time at which the mean changed is estimated by tracking the line segment leading to the breakout back to the last occasion on which it lay on the horizontal axis. Thus, in the case of systolic blood pressure, the breakout occurs on day 30 (16 days after the start of lisinopril treatment) and the shift in pressure is estimated to have occurred on day 22 (8 days after lisinopril treatment started). © Halberg.



Figure 17. Control charts of decimated time series based on single daily measurements (top, left and right; bottom left) and of a mean of three daily measurements (bottom right) of blood pressure and heart rate are shown with shaded decision intervals for a shift in location index by 1 standard deviation. Whereas single daily measurements are insufficient to reach a consistent decision whether lisinopril was effective, such a conclusion can be made when three daily measurements are used. © Halberg.



Figure 18. In many circumstances the data are collected or are amenable to being collected over time but the information is discarded after a mere visual inspection of a monitor's recording. Adding chronobiologic analytical procedures for the on-line processing and interpretation of the data would provide individual reference standards for rhythms with lower and lower frequencies while also providing continued check-ups capable of detecting the earliest rhythm alterations indicative of a heightened risk and thus enabling the prompt institution of treatment when indicated. (The analogy that a single 24-hour cycle from a circadian viewpoint corresponds to the radial pulse based on a single heartbeat [42] has now been shared by others [43].) © Halberg.



Scheme 1. Abstract demonstration of the amplitude on top. This endpoint is a measure of extent of predictable change, recurring with a certain period; when the period is 24 hours, the amplitude can be described as circadian. Note that the total predictable change within 24 hours is the double amplitude. In the also-abstract bottom part of the graph, an excessive amplitude is demonstrated, without any change in overall mean. © Halberg.



Table 1.