There is mounting evidence that the parameters of rhythms can be used as harbingers of risk before the onset of overt disease. This is the case in particular for blood pressure, or BP. Currently, only single or a few measurements of systolic blood pressure, or SBP, and diastolic blood pressure, or DBP, are taken, usually in the physician’s office. The distinction between normotension and hypertension is made on the basis of whether SBP/DBP values are below or above 140/90 millimeters of mercury, or mm Hg. Since BP is highly variable, single measurements provide only snapshots that may not be at all representative of the patient’s BP at other times. The availability of ambulatory monitors that can be programmed to inflate an arm cuff at regular intervals throughout the day and night makes it feasible to obtain much more reliable information—not only about the mean value during each day, but also about the dynamics of BP and heart rate.

The diagnosis of ‘hypertension’ cannot be safely ruled in or out in a given person on the basis of single in-office measurements of blood pressure.

An elevated BP increases the risk of vascular complications, as established by single measurements on groups of people.^1–4 Even if the measurement taken in the physician’s office is obtained by a trained professional with a well-calibrated instrument, it is representative only of the patient’s BP at that precise moment, which may differ drastically from values at other times. Since a person’s BP can vary substantially, a single measurement will not represent his or her usual BP dynamics accurately.⁵ When single measurements of BP are used as a reference, results with respect to the effect of BP on the risk of death will be biased.^6,7

With such qualifications, a few BP measurements are used for many people, but they may not substitute for many chronobiologically interpreted measurements needed for a reliable assessment of the given person’s BP and heart rate status.⁸ The diagnosis of "hypertension" cannot be safely ruled in or out in a given person on the basis of single in-office measurements of BP.^8–10 In view of the large inter- and intraindividual variability in BP, not only from hour to hour but also from day to day and as a function of age, it seems desirable to estimate in each case the parameters of variability in BP:

– the MESOR;

– the circadian amplitude, a measure of the extent of daily change;

– the circadian acrophase, a measure of the timing of overall high values occurring each day.

These parameters are approximated by the fit of, for example, a 24-hour cosine curve, along with nonparametric chronobiological endpoints.^9,10 (The fit of an additional seven-day cosine curve would provide information regarding changes along the scale of the week as well.)

Nonparametric endpoints include the hyper-baric index, computed by numerical integration of the area between the time-varying upper 95 percent limit of a prediction interval and the subject’s BP curve, where it exceeds this limit (computed from data of normal peers of the same sex and a similar age).¹⁰

Parametrically, one can detect a separate BP-related risk of stroke and nephropathy, as well as myocardial infarction or retinopathy—a systolic and/or diastolic pressure that overswings along the 24-hour scale, a phenomenon called circadian hyper-amplitude-tension, or CHAT.^9,10 CHAT technically is defined as a circadian amplitude larger than the upper (95 percent) limit of the 90 percent prediction interval calculated for people of a given sex and age in a data bank established at the University of Minnesota in the 1950s. The data bank on ambulatory monitoring that has accumulated during the last decade has been extended into an international project on the biosphere and the cosmos, known as BIOCOS.¹¹ This project involves the monitoring of physiological functions, such as BP and heart rate, in different geographic locations (Europe, Asia and North America), concomitantly with physical variables to determine ethnic or environmental influences.

A seven-day/24-hour monitoring of BP and heart rate is needed for a diagnosis of an above-threshold variability of BP (CHAT, as described above) that can be complicated by a below-threshold variability in heart rate.⁹ This monitoring also serves to detect an overall elevation in BP, the proven etiopathogenic factor. Indeed, a high BP can be detected much more reliably with the seven-day/24-hour approach, since all of 45 conventionally diagnosed hypertensive patients in an unpublished independent study had normotensive diastolic values during office hours (Y. Watanabe, M.D., Ph.D., and colleagues, unpublished data, 1997) (Figure 1). Figure 2 demonstrates the merit of weeklong automatic monitoring. In other words, normotensive patients can have high pressures during office hours and hence may be placed on medication they do not need.⁹ Even a 24-hour ambulatory BP average can be high for several days and then acceptable for months or years of ambulatory monitoring thereafter, as illustrated elsewhere in single cases.⁹ For instance, a 33-year-old man who monitored himself for 30 days had a rhythm-adjusted 24-hour average of SBP that varied between less than 120 and greater than 130 mm Hg, suggesting that a week’s monitoring may not always suffice for a reliable diagnosis.

View larger version (38K): [in this window] [in a new window]   Figure 1. Distribution of 45 patients conventionally diagnosed as hypertensive according to the percentage of systolic blood pressure, or SBP, and diastolic blood pressure, or DBP, below World Health Organization, or WHO, limits of 140/90 millimeters of mercury observed during daytime hours (8 a.m. to 5 p.m.). Reproduced with permission of Franz Halberg, M.D.

View larger version (40K): [in this window] [in a new window]   Figure 2. Distribution of 78 patients conventionally diagnosed as normotensive according to the percentage of systolic blood pressure, or SBP, and diastolic blood pressure, or DBP, above World Health Organization, or WHO, limits of 140/90 millimeters of mercury observed during daytime hours (8 a.m. to 5 p.m.). Reproduced with permission of Franz Halberg, M.D.

Blood pressure screening is of interest not only to maximize patient safety during dental visits, but also because of the dentist’s concern for preventive health care beyond the teeth, periodontium and oral lesions.

Screening is of interest not only to maximize patient safety during dental visits, but also because of the dentist’s concern for preventive health care beyond the teeth, periodontium and oral lesions. Our design of a study on which we published a report in JADA in April 1998¹⁴ relied on extensive BP monitoring, which made this follow-up possible.

	METHODS AND MATERIALS

TOP ABSTRACT METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

 
The original study, conducted in 1990 and 1991, involved 24 consecutively seen patients. According to conventional criteria, they presumably were normotensive.¹⁴ Each subject came to the dental office for three appointments; the first two involved deep scaling of the periodontium under local xylocaine anesthetic, with or without epinephrine, and the third involved periodontal surgery for open débridement and apically repositioned flaps.

The patients carried an ambulatory BP monitor (ABPM-630, Colin Medical Instruments, Komaki, Japan), first for four days, then for two days and finally for three days. Data were obtained at 15-minute intervals, except for the sessions in the dental chair that involved measurements at five-minute intervals for spans of approximately 120 minutes. Each BP profile bracketed a dental appointment, which was made at a few-weeks’ interval from the one before or after.

To each data series in a given profile, we fitted a 24-hour cosine curve by least squares. We compared estimates of the MESOR, circadian amplitude and acrophase of SBP and DBP, and heart rate for each subject with the 90 percent prediction limits established previously for clinically healthy peers matched by sex and age.

About seven years after the dental patients underwent this monitoring, we made an attempt to recall them to find out about their general health status. We aimed to answer one central question: how well does a single measurement of BP stand up to measurements taken over nine days of monitoring?

	RESULTS

TOP ABSTRACT METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Twenty-three of the 24 patients enrolled in the study completed the entire nine-day measurement schedule. Whereas all patients presumably were normotensive, our chronobiological interpretation of the data collected automatically around the clock for several days detected abnormality. Therefore, patients thought to be normotensive, as gauged by office measurements made casually and interpreted in terms of time-unqualified reference values, in fact can be chronobiologically deviant.

In a total of nine days of monitoring, only 10 of the 23 patients conventionally diagnosed as normotensive actually were free of BP deviation in all three sessions. Thirteen (56.5 percent) had some chronobiologically documented abnormality. Eleven of those (47.8 percent of the total 23) had BP excess above 50 mm Hg x h during 24 hours during one or several of the sessions, and three of them (13 percent) had excess in all three sessions and CHAT in at least one session. Of 20 patients with abnormality absent in at least one session, one had no excess in the first four-day session but had excess in the subsequent two- and three-day sessions, and one had an elevated SBP MESOR during the first and third sessions but not during the second. Among the remaining patients, abnormalities were limited to one session only.

First follow-up. After seven years, we attempted to contact all patients who had participated in the study described earlier. Eight patients of the original 23 responded by phone to the follow-up. One of these eight participants reached had experienced a myocardial infarction. This male participant also was the only one among the eight patients recalled who had an elevated MESOR of SBP in each session, as well as systolic CHAT in two sessions and diastolic CHAT in one session. Thus, in each session, he had some abnormality. While he was a non-smoker, he was the oldest patient. This additional risk factor should be considered in the light of the finding in a six-year prospective study that the risk associated with CHAT was numerically larger and independent of the risk associated with an advanced age.¹⁵ When we used a Fisher exact test, we found that the results for the eight patients who were reached first for follow-up were not statistically significant enough to serve as documentation of an association with an untoward event.

Second follow-up. Later, four more patients were located and contacted, boosting the total of previously monitored subjects to 12. Among these four, one subject had had a chronobiological abnormality in each of the three monitoring sessions. This subject reported that since the study, he had had coronary artery bypass surgery, interpreted for our purposes as an indication of advanced myocardial disease. The other three, none of whom had had abnormalities in any of the three sessions, had not experienced an untoward event since the study. Thus, the only two patients of 12 who had had abnormalities in all three sessions also were the only ones who had experienced morbid events.

With the summary of all 12 recalled patients, the difference in outcome between patients with vs. without chronobiological abnormality in all three sessions during nine days of monitoring had become statistically significant (P = .015 by Fisher exact test) (Figure 3).

View larger version (34K): [in this window] [in a new window]   Figure 3. A comparison of the incidence of adverse outcomes in patients with blood pressure, or BP, abnormality in all three sessions vs. that in patients with abnormality in two, one or none of the sessions, for each of the two recalls (I and II, left and middle), using circadian hyper-amplitude-tension, or CHAT, and/or diastolic blood pressure excess (elevated hyperbaric index above 50 millimeters of mercury x hour during 24 hours). P values were derived using Fisher exact test. N: Number. CI: Confidence interval. No untoward event occurred among patients with BP abnormality in two, one or none of the sessions, whereas both patients who experienced a vascular problem had been diagnosed with CHAT and/or other abnormalities in all three sessions. The prognostic value of a consistently (in three of three sessions) deviant chronobiological assessment is illustrated by the relative risk obtained as the ratio of incidences between the two groups being compared (right). Reproduced with permission of Franz Halberg, M.D.

	DISCUSSION

TOP ABSTRACT METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

Because of low BP values during sleep by night, it has been emphasized by others²² and ourselves²³ that the upper limit of acceptability for around-the-clock automatic measurement series should be lower than 140/90 mm Hg, a limit specified by the World Health Organization, or WHO, for measurements in the care provider’s office. We use sex, age and, when possible (in 19 of 23), ethnicity-specified limits here.

When conventionally (nonchronobiologically) defined abnormality was the criterion of high risk, there were three false-positives. The chrono-biological approach on this admittedly small sample identified as abnormal only the two patients who both had actual events; in other words, there were no false-positives.

In view of variability, shown in Figures 1 and 2 (and elsewhere¹⁴), single office measurements of BP do not suffice. The fact that a single BP measurement is below WHO’s upper limit of acceptability (140/90 mm Hg) is no reason to tell the person whose BP was measured that nothing is wrong and he or she is merely to come back in two years (as is WHO’s current recommendation).^2,3 When the single measurements are interpreted in the light of data series, it becomes apparent that they are no more than a snapshot of a roller coaster.⁹ Even the current gold standard of a 24-hour BP profile can be unreliable to the point that it has been likened to "flying blind."⁹ As a result, people who may not need treatment receive drugs that they do not need, while others who are at a high risk of experiencing vascular complications remain untreated.¹⁰ It is important scientifically and clinically for the health of the patient to avoid false-positive and false-negative diagnoses as much as possible. This is true especially in light of the significant adverse side effects of some cardiac medications. In view of the time spent by patients in the dental chair during an appointment, dentists are well-placed to explain to them the advantages of chronobiological BP monitoring.^14,19–21 We here add evidence showing what dentists can do with ambulatory BP monitoring for people diagnosed as normotensive by conventional criteria.

If the variability of blood pressure is too large, even if the mean is acceptable, there is a great risk of stroke and other vascular diseases, some of them costly cripplers.

The dentist easily can provide guidance about the merit of using the extent of circadian swing to diagnose a high risk of stroke, via CHAT, which is a diagnosis that cannot be based on casual measurements. He or she has the opportunity to persuade the patient to monitor BP for longer than a day—for instance, for a week, at first—so that any diagnosis is more reliable than that based on a profile built on the snapshot taken on a single day.

In the United States, the average fee for a 24-hour profile is $200 (usually paid by the patient, not the insurer, at least in our experience). This makes the service inaccessible to most people. The cost of seven-day monitoring, however, may be a small price to pay for reducing the uncertainty of a diagnosis that, in our current state of lack of etiologic knowledge, may prompt treatment to continue for a long time. The dentist can describe for the patient the many advantages of monitoring and analyzing, rather than spot-checking. The dentist also can explain that the diagnosis is much more reliable, if one asks not only whether the location of the BP along a scale in mm Hg is or is not acceptable, but also whether the variability of measurements is acceptable, too low or too high (too low a variability in heart rate is a prognostic factor for coronary artery disease). The reason for the broader inquiry cannot be overemphasized. If the variability of BP is too large, even if the mean is acceptable, there is a great risk of stroke and other vascular diseases, some of them costly cripplers.⁹

Our results support the recommendation that automatic or manually activated monitors belong in the dental office. This suggestion can be extended to the office of any other health care provider as well—nurses, pharmacists, physicians. Eventually, monitoring should become a matter of self-help, as is self-care by people with diabetes—with the important distinction that abnormalities can be detected within the current physiological range, before the onset of overt disease and the appearance of symptoms, thus allowing for prophylactic intervention. If the dentist provides the instrumentation the patient needs for self-monitoring, presenting it as "research" in everyday life, the market would open because of a larger demand. The cost of ambulatory monitoring then could be greatly reduced and the procedure would become affordable.

In principle, the evaluation of a person’s BP, notably in the case of patients with BP disorder, "should include the estimation of the ‘true’ BP (the average over prolonged periods of time)"; by the same token, ideally "the absolute level of risk" should be assessed.²⁴ This recommendation properly places emphasis on risk, which is the more important priority, since risk is best assessed by the variability (the time structure or chronome of BP⁹) rather than merely by its average over prolonged spans of time. Over time, there is no "true" BP, but rather a time structure or chronome of this variable. In studying the variability systematically, whether by self-measurement with manually activated or with ambulatorily functional instrumentation, one obtains a better average if data are evaluated chronobiologically. Usually the MESOR is more accurate than the arithmetic mean in data that are available only at unequal intervals (more densely during waking than by night, as in the case of self-measurements). The MESOR is more precise as a rule in the case of equidistant data that can be obtained easily with ambulatory instrumentation; it then usually has a smaller standard deviation than the arithmetic mean.

Self-monitoring is cost-effective as well as useful, a point made by the Kaiser Foundation,²⁵ in keeping with recommendations made repeatedly by some of us.^{9,14,16,19–21,26} Dentist-guided self-monitoring, preferably with instruments that can be worn while a patient goes about daily activities, is most effective when accompanied by a chronobiological analysis that offers not only an improved mean, but also the additional endpoints describing dynamic characteristics of change in BP or heart rate.^27–30 Endpoints of variability are the more interesting and useful, since an excessive circadian amplitude of BP, notably when it is accompanied by a below-threshold variability in heart rate, is the greatest risk factor of all tested. This is documented in a six-year prospective study of 297 patients who were monitored for two days at 15-minute intervals and followed up at six-month intervals for six years.^9,27 Among 253 patients who had an acceptable circadian BP amplitude and an acceptable 24-hour standard deviation of heart rate, only 20 (8 percent) experienced a morbid event. By comparison, seven of 20 (35 percent) patients diagnosed with diastolic CHAT and eight of 19 (42 percent) patients with a reduced heart rate variability had morbid events. Among the five patients who had both conditions, four (80 percent) had morbid events.

In our most recent follow-up of 63 patients whom we studied 28 years ago, we tracked down 21 patients. In the original study, manual BP measurements had been obtained at two- to three-hour intervals during waking for two days.³¹ Among the 21 patients, only one of 10 still alive had systolic or diastolic CHAT, as compared with seven of 11 who died (Figure 4). The difference (P< .01 by ² test) further attests to the value of detecting CHAT before providing dental treatment—for example, by instructions given in the dental chair and with instrumentation provided by the dentist. (Authors’ note: Colleagues interested participating in the BIOCOS project can make arrangements to obtain instruments at a cost highly reduced—by 90 percent—as well as analyses of specifically formatted data in the light of an accumulating database, qualified by sex and age, from the Chronobiology Center at the University of Minnesota. Interested readers should contact Dr. Germaine Cornélissen at "corne001{at}tc.umn.edu".)

View larger version (57K): [in this window] [in a new window]   Figure 4. There is a statistically significant difference in the incidence of circadian hyper-amplitude-tension, or CHAT, between those who lived and those who died, within the 28 years that elapsed between the original monitoring (by manual measurements of blood pressure for two days) and the recall of the patients. Reproduced with permission of Franz Halberg, M.D.

	CONCLUSION

TOP ABSTRACT METHODS AND MATERIALS RESULTS DISCUSSION CONCLUSION REFERENCES

Thus, CHAT, a condition defined as an excessive circadian BP amplitude (above the upper 95 percent prediction limit of healthy peers), is associated with an increase in the risk of a cerebral ischemic event by 720 percent. A reduced 24-hour standard deviation of heart rate (below the lower 5 percent prediction limit of healthy peers) likewise is associated with an increase in the risk of experiencing coronary artery disease by 550 percent. Such large increases in the risk of morbid events are present even in MESOR-normotensive patients who would not receive treatment according to current standards. The increase in risk associated with CHAT is so large that even with the small sample size available in our study, the relative risk associated with a consistently deviant chronobiological assessment is larger than unity (representing equal risk). The nonoverlap of unity by the 95 percent confidence interval indicates that the corresponding increase in risk is statistically significant.

Because dental procedures may place some compromised patients at cardiovascular risk, this information concerns the profession as a whole. Dentists usually can spend more time with their patients than can physicians; therefore, they could take the lead in offering a chronobiological BP monitoring and in educating their patients in the merits of such monitoring. By extending use of ambulatory BP monitoring to all patients, rather than restricting it to special cases as is the current practice, dentists could help decrease the cost of monitoring dramatically, and thereby render the procedure affordable for all.